Hybrid Regimes within Democracies: Fiscal Federalism and Subnational Rentier States 1316510735, 9781316510735

Table of contents :
Cover
Half Title
Title Page
Imprints Page
Dedication
Contents
List of Figures
List of Tables
Preface and Acknowledgments
List of Abbreviations
Introduction
I.1 Getting on the Agenda: Comparative Politics and Subnational Regimes
I.2 Description: Objective and Subjective Operationalizations of Subnational Democracy
I.3 Explanation: A Rentier Theory of Subnational Democracy
I.4 Democracy, Federalism, and Subnational Politics in Argentina
I.5 Plan of the Book
Part I. Description: The Anatomy and Evolution of Subnational Regimes
1 Defining and Measuring Subnational Regimes
1.1 The Background Concept: Democracy
1.1.1 Level or Quality of Democracy?
1.2 The Systematized Concept: Liberal Representative Democracy
1.2.1 The Other End: Authoritarianism or Hybrid Regimes?
1.2.2 The Dimensions and Subdimensions of the Concept
1.2.2.1 The Democratic Dimension: Contested, Inclusive, and Effective Elections
1.2.2.2 The Liberal Dimension: Institutional Constraints and Individual Rights
1.2.2.3 Components and Subcomponents of Democracy
1.3 Indicators: Objective and Subjective Measures of Democracy
1.4 Aggregation: From Indicators to Indices of Subnational Regimes
1.5 Conclusion
2 The Subnational Democracy Index: Trends in Provincial Regimes (1983−2015)
2.1 Objective Indicators: The Subnational Democracy Index
2.2 Results of the Subnational Democracy Index
2.2.1 Subnational Regime Variance: Cross-sectional and Temporal Components
2.3 Democratic and Hybrid Provinces: Qualitative Evidence on Regime Differences
2.4 Conclusion
3 Expert Survey Evidence: The Many Dimensions of Subnational Democracy
3.1 Subjective Indicators: The Survey of Experts on Provincial Politics
3.1.1 Aggregation: From Individual Responses to Provincial Indices
3.2 Results of the Survey of Experts on Provincial Politics
3.2.1 Two Dimensions of Subnational Regimes: Incumbency Advantage and Repression
3.2.2 First- and Second-level Indices of Subnational Democracy: National Level
3.2.3 First- and Second-level Indices of Subnational Democracy: All Provinces
3.2.4 A Look at the National Distribution of Individual Items
3.3 Comparing Objective and Subjective Measures of Subnational Democracy
3.4 Conclusion
Part II. Explanation: The Causes of Subnational Regimes
4 On the Rentier Effects of Fiscal Federalism on Subnational Regimes
4.1 Theories of Subnational Democracy
4.2 State-Society Balance, Fiscal, and Rentier Theories of Democracy
4.2.1 Democracy as the Effect of State-Society Balance
4.2.2 Democracy as the Effect of Fiscal Bargains
4.2.3 Authoritarianism as the Effect of Natural Resource (and Other) Rents
4.2.4 Statism: Where Balance, Fiscal and Rentier Theories Meet
4.3 Climbing up the Ladder of Abstraction: From “Resource Rents” to “Fiscal Rents”
4.4 Climbing Down to the Subnational Level: From “Rentier States” to “Rentier Regions”
4.5 Conceptualizing Fiscal Rents
4.6 The Subnational Rentier State: Fiscal Federalism Rents as a Special Case of Fiscal Rents
4.7 Are Fiscal Federalism Rents more Archetypical than Resource Rents?
4.8 Conclusion
5 Fiscal Federalism, Subnational Rentierism, and Hybrid Provincial Regimes in Argentina
5.1 The Argument: Fiscal Federalism Rents against Subnational Democracy
5.2 Fiscal Federalism as a Source of Subnational Rentierism
5.3 From Fiscal Federalism Rents to Hybrid Regimes: Exploring Causal Mechanisms
5.4 Conclusion
6 The Determinants of Provincial Regimes in Argentina
6.1 Test 1: Cross-Section–Times-Series Data on the subnational democracy index
6.1.1 Operationalization
6.1.2 Estimation
6.1.3 Results
6.1.4 Results with Alternative Measures of Rentierism
6.1.5 Endogeneity: Spurious and Simultaneous Causality
6.1.6 Detecting Deviant Observations: Residual Analysis
6.2 Test 2: Cross-­Sectional Data on the Subjective Indices of Subnational Regimes
6.2.1 Operationalization
6.2.2 Estimation
6.2.3 Results
6.2.3.1 Results for the Two Second-Level Indices
6.2.3.2 Results for the Ten First-Level Indices
6.2.4 Endogeneity: Simultaneous Causality
6.3 Causal Mechanisms: The Evidence
6.4 Conclusion
Part III. Comparison: Subnational Regimes Around the World
7 A Comparative Perspective: Levels of Subnational Democracy in Seven Federations and One Unitary Country
7.1 A Simple (and Rough) Alternative: The (Subnational) ACLP Index
7.2 A New Objective Measure: The Comparative Subnational Democracy Index
7.3 The Comparative Subnational Democracy Index Applied to Eight Nations
7.4 On the Cross-Country Validity of the CSDI
7.5 Conclusion
Conclusion
Policy Implications
Appendix A Methodological Design of the Survey of Experts on Provincial Politics (SEPP)
Appendix B From Survey Items to Indices of Subnational Democracy
Appendix C Measurement Error Estimates
References
Index

Citation preview

Hybrid Regimes within Democracies From the racially segregated “Jim Crow” US South to the many electoral but hardly democratic local regimes in Argentina and other federal democracies, the political rights of citizens around the world are often curtailed by powerful subnational rulers. Hybrid Regimes within Democracies presents the first comprehensive study of democracy and authoritarianism in all the subnational units of a federation. The book focuses on Argentina, but also contains a comparative chapter that considers seven other countries including Germany, Mexico, and the United States. The in-depth and multidimensional description of subnational regimes in all Argentine provinces is complemented with an innovative explanation for the large differences between those that are democratic and those that are “hybrid” – complex combinations of democratic and authoritarian elements. Putting forward and testing an original theory of subnational democracy, Gervasoni extends the rentier-state explanatory logic from resource rents to the more general concept of “fiscal rents,” including “fiscal federalism rents,” and from the national to the subnational level. Carlos Gervasoni is Associate Professor of Political Science and International Studies at Universidad Torcuato Di Tella (Buenos Aires, Argentina). His articles have appeared in journals such as Comparative Political Studies, Democratization, Party Politics, Política y Gobierno, and World Politics. He is a regional manager for the Varieties of Democracy project and a member of the 2015 Argentine Panel Election Study research team. His research on subnational regimes has been supported by a National Science Foundation award.

Hybrid Regimes within Democracies Fiscal Federalism and Subnational Rentier States

CARLOS GERVASONI Universidad Torcuato Di Tella

University Printing House, Cambridge CB2 8BS, United Kingdom One Liberty Plaza, 20th Floor, New York, NY 10006, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia 314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre, New Delhi – 110025, India 79 Anson Road, #06-04/06, Singapore 079906 Cambridge University Press is part of the University of Cambridge. It furthers the University’s mission by disseminating knowledge in the pursuit of education, learning, and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781316510735 DOI: 10.1017/9781108590679 © Carlos Gervasoni 2018 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2018 Printed in the United States of America by Sheridan Books, Inc. A catalogue record for this publication is available from the British Library. Library of Congress Cataloging-in-Publication Data ISBN 978-1-316-51073-5 Hardback Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

To the memory of my father Carlos Rafael Gervasoni. To my mother Kety. To Sybil and our wonderful son Federico.

Contents

List of Figures List of Tables Preface and Acknowledgments List of Abbreviations Introduction part i.  description: the anatomy and evolution of subnational regimes 1 Defining and Measuring Subnational Regimes 2 The Subnational Democracy Index: Trends in Provincial Regimes (1983−2015) 3 Expert Survey Evidence: The Many Dimensions of Subnational Democracy part ii.  explanation: the causes of subnational regimes 4 On the Rentier Effects of Fiscal Federalism on Subnational Regimes 5 Fiscal Federalism, Subnational Rentierism, and Hybrid Provincial Regimes in Argentina 6 The Determinants of Provincial Regimes in Argentina part iii.  comparison: subnational regimes around the world 7 A Comparative Perspective: Levels of Subnational Democracy in Seven Federations and One Unitary Country

page ix xi xiii xxi 1

23 42 67

111 137 155

215

vii

Contents

viii

Conclusion Appendix A  Methodological Design of the Survey of Experts on Provincial Politics (SEPP) Appendix B  From Survey Items to Indices of Subnational Democracy Appendix C  Measurement Error Estimates References Index

238 251 255 258 261 281

Figures

  I.1 Overall evaluations of subnational democracy in Argentina (2003−7) page 7   I.2 Subnational elections in nineteen large federations (2014) 8   I.3 Argentina’s provinces and their level of democracy (1983–2015) 16   1.1 Genus, differentia, dimensions, and subdimensions of democracy 33   1.2 A graphical representation of national and subnational regime variance and its measurement 36   2.1 Average Subnational Democracy Index (1983−2015) by province 53   3.1 Average of seventeen first-level indices of subnational democracy (2003−7) by province 79   3.2 Scatterplot of Repression factor and Incumbency Advantage factor 85   3.3 Boxplot of seventeen first-level indices (ordered from lowest to highest median) 87   3.4 Scatterplot of indices’ means and standard deviations by subdimension 88   3.5 Provincial scores on the Fair Elections index and Campaign Advantage index 90   3.6 Provincial scores on Media Bias index and Pluralistic Media index 91   3.7 Provincial scores on the Hard and Soft Media Control indices 92   3.8 Provincial scores on the Judicial and Legislative Control indices 93   3.9 Provincial scores on the Totalitarian Control index and Horizontal Accountability index 95 3.10 Provincial scores on the Police Repression index and Government Discrimination index 96

ix

x

List of Figures

3.11 Provincial scores on the Media Independence index and Punish Opponents index 3.12 Scatterplot of item means and standard deviations by subdimension 5.1 A stylized model of fiscal federalism rents 5.2 The determinants of federal transfers per capita (2003−7) 6.1 The effect of Federal Transfers pc on subnational democracy conditional on logged GGP per capita 6.2 Average combined residual per province 6.3 Scatterplots of the Punish Opponents index on measures of rentierism and development 6.4 Scatterplots of the Media Independence index on measures of rentierism and development 7.1 Alternative mappings of executive incumbent vote (EIV) to estimated levels of democracy 7.2 Boxplot of Comparative Subnational Democracy Index in eight selected countries (from circa 1983 to 2015) 7.3 Dotplot of least democratic regions by country (observation with CSDI < 30)

97 100 141 145 170 177 182 183 224 231 232

Tables

1.1 Disaggregation of the concept of subnational democracy page 34 1.2 Advantages and disadvantages of objective and subjective indicators of democracy 38 2.1 Provincial democracies and dictatorships according to the Alvarez et al. (1996) index (as of 2015) 50 2.2 The Subnational Democracy Index: Summary statistics and temporal trends by province (from less to more democratic), 1983−2015 52 2.3 The Subnational Democracy Index: Summary statistics over time 56 3.1 SEPP-based indices of subnational democracy 75 3.2 Factor analysis of seventeen first-level indices of subnational democracy; Rotated factor loadings and unique variances 83 3.3 Provincial scores in all indices 99 3.4 Correlations between the objective Subnational Democracy Index and twenty-two subjective indices of democracy 105 4.1 Typology of definitional attributes of the concept of rent 130 5.1 Economic, demographic, and fiscal indicators for Argentina’s provinces 148 5.2 The determinants of public employees and own taxes (random effects models) 153 6.1 Summary statistics for dependent and independent variables 163 6.2 The determinants of subnational democracy (I), 1983−2007 (random-effects models) 164 6.3 The determinants of subnational democracy (II), 1983−2007 (random-effects models) 168

xi

xii

List of Tables

6.4 The interactive effect of fiscal rents and development, 1983−2007 (random-effects models) 6.5 The determinants of subnational democracy, 1983−2007 Alternative measures of rentierism (random-effects models) 6.6 Summary statistics for dependent and independent variables 6.7 The determinants of SEPP second-level indices (OLS models) 6.8 The determinants of SEPP second-level indices. Alternative measures of rentierism (OLS models) 6.9 Regression of ten SEPP first-level indices of democracy on Federal Transfers and Resource Rents (OLS models) 6.10 Regression of ten SEPP first-level indices of democracy on Fiscal&Resource Rents (OLS models) 6.11 Regression of ten SEPP first-level indices of democracy on Rentierism (OLS models) 7.1 Number, percentage, and name of nondemocratic subnational units (according to the ACLP Index) in eight selected countries (from circa 1983 to 2015) 7.2 Summary statistics for Comparative Subnational Democracy Index (CSDI) in eight selected countries (from circa 1983 to 2015)

169 172 180 184 185 188 189 190

217

230

Preface and Acknowledgments

This book is a way of expressing frustration about the politics of Argentina, my country. Following the principle that one can use normative considerations for choosing a research topic, though not for drawing empirical conclusions about it, I decided to study the places from which came those, in my opinion, deplorable presidents who dominated Argentine politics in the 1990s and 2000s – Peronists Carlos Menem and Néstor Kirchner. Their seemingly sharp ideological differences – the former a “newborn” neoconservative, the latter a populist and alleged leftist – may have hidden for some the many similarities between them. Of all of these – both frequently overstepped constitutional rules, both filled important government positions with unqualified relatives and friends, both led utterly corrupt administrations, both were rich, both were Peronist – there is one that escapes most analyses: before becoming presidents, they were the hegemonic governors of provinces that – as documented in this book – are among the country’s least democratic. La Rioja and Santa Cruz seem to have served Menem and Kirchner, respectively, as training camps from where to launch their bid for the presidency, which they exercised in ways comparable to the way they ruled those provinces. Their fellow satraps, such as Vicente and Ramón Saadi (Catamarca), Gildo Insfrán (Formosa), or Carlos Juárez and Gerardo Zamora (Santiago del Estero) likely had, or have, similar dreams. That the governors of two of the demographically smallest Argentine provinces (together they account for 1.5 percent of the country’s population) were able to prevail over those of Buenos Aires, Córdoba, Santa Fe, or Mendoza – which contain 60 percent of the country’s population – may be more than a coincidence. In fact, a third Peronist governor of a small and hardly democratic province – Adolfo Rodríguez Saá of San Luis – also occupied the presidency briefly during the political crisis of 2001–2002. There seems to be a political logic at work: “executive malapportionment” (each province, small or large, has one governor) is compounded by the fact that small provinces are heavily subsidized by Argentina’s fiscal federalism, and by the fact that these subsidies xiii

xiv

Preface and Acknowledgments

allow their rulers to establish hegemonic regimes which, in turn, are functional for their national political careers. The governors of Buenos Aires or Santa Fe have to make do with comparatively little national fiscal support and face strong provincial challengers; the governors of La Rioja and Santa Cruz, on the other hand, can rely on plentiful “free” fiscal resources – what I call “fiscal federalism rents” – and have little to worry about in terms of the provincial opposition. In summary, if the causal logic uncovered in the pages ahead is correct, we Argentines have given ourselves institutions of fiscal federalism that subsidize the emergence and survival of semi-authoritarian provincial bosses that have good chances of becoming presidents, and that typically bring their undemocratic ways with them on their trips from La Rioja, San Luis, or Santa Cruz to the capital city of Buenos Aires. The concepts, theories, and evidence I present help us better understand the type of provincial regimes they led (and the more democratic regimes in other provinces) and the causes that account for the coexistence of more and less democratic provinces under the same national regime. Beyond my normative inclinations, this book is an attempt to remediate what Robert Dahl called a “grave omission” of his influential Polyarchy, that is, the neglect of subnational regimes. I had the good fortune to do my doctoral research under one of Dahl’s Yale students, Michael Coppedge. A superb advisor, Michael helped me turn some rudimentary ideas about provincial regimes in Argentina into the set of theoretically grounded and methodologically sound (or so I hope) descriptive and causal inferences that structure this book. If its motivation has been largely normative, its execution has been as scientific as the capacities of its author permitted. Dahl and the other founding parents of the study of political regimes (such as Juan Linz, Michael’s dissertation advisor) were morally appalled by the realities of twentieth-century authoritarianisms and totalitarianisms, and responded by launching a scientific research program that has proved fruitful, cumulative, and influential. I take pride in working within it, and in helping fill gaps that the founders alerted us to. The realities I react to, fortunately, are far from the horrors of Nazi Germany or the Soviet Union that they had in mind. But as a liberal democrat who is always wary of the state’s potential oppressive deviations, and as an Argentine who has seen his own state murder its own citizens (first under the Juan Perón−Isabel Perón [1973–1976] administration, when a government-sponsored paramilitary force [the Triple A, or Alianza Anticomunista Argentina] committed hundreds of crimes, and then, on a much larger scale, under the 1976−1983 military dictatorship) I do not underestimate the potential threats to freedom that emerge even from elected governments. However undemocratic the administrations of governors such as Menem or Kirchner may have been, they were very far from these horrors. Many citizens in their provinces, however, had their political rights restricted, and all the country’s citizens started to live in a less democratic national regime when Menem and Kirchner became presidents. Menem,

Preface and Acknowledgments

xv

for example, quickly packed the Supreme Court with a group of loyal and often-unqualified cronies that made a joke of judicial checks and balances. Kirchner, among other things, targeted one of the most basic pillars of democracy – the critical media. No wonder: there are hardly any independent judges in La Rioja and hardly any independent media outlets in Santa Cruz. The first instinct of these leaders was to reproduce at the national level the hegemonic realities they had become used to in their provinces. They sometimes succeeded and sometimes failed, but they never ceased to be a danger for the political rights of Argentines. Democracy is at times blatantly and bloodily prevented or suppressed, by the Hitlers, Stalins, and Videlas of this world, but at other times it is less cruelly undermined from within, for example by the Kirchners, Menems, and Wallaces of the subnational world. I wrote this book with the Enlightenment’s faith in rational and empirical knowledge as a tool for social improvement. I do not ignore that, in the realm of policymaking, political interests more often than not trump scientific argument, but trust that a clearer understanding of politics is a precondition for positive change. The book presents a description of subnational regimes in Argentina (including a comparison with several other countries around the world) and an anatomy of the typical undemocratic practices found in them. It also presents a theory that identifies a very clear culprit for low levels of subnational democracy. In showing that provinces such as Formosa, La Rioja, San Luis, Santa Cruz, and Santiago del Estero are much less democratic than others such as Buenos Aires, Córdoba, Entre Ríos, Mendoza, or Santa Fe, and in arguing that the difference lies (to a large extent) in the former’s access to plentiful federal subsidies (or fiscal federalism rents), this book hopes to help Argentines understand our own country, help scholars understand subnational regimes, and help make those regimes more democratic. The following pages would have never been written without the great personal and intellectual support of a wonderful group of professors at the Department of Political Science and the Kellogg Institute for International Studies at the University of Notre Dame, especially Michael Coppedge, Frances Hagopian, and Scott Mainwaring. I have seen them teach great, well-prepared courses, work hard at advising students, make an effort to be as fair as humanly possible in their decisions, and be truly humble regarding their own academic achievements. In making tough decisions about advisors during my doctorate, I applied the rule of always working with people who are professionally and morally good. The rule worked as wonderfully as ever. Along with Benjamin Radcliff, who provided the outside view of an Americanist (and non-Latinoamericanist comparativist), the rigor and insightfulness of an amazingly published social scientist, and the normative commitment of a good heart, Michael, Fran, and Scott supported my research beyond duty. I run out of words to express my appreciation and thankfulness toward them. I owe an immense intellectual debt to Michael Coppedge. His wonderful seminars on the causes, consequences, and measurement of democracy,

xvi

Preface and Acknowledgments

his constant disposition to talk during long hours, his wise suggestions, and his personal support were critical in producing the research I present below. Inadvertently, he taught me a lesson or two about how to combine rigorous academic work with an open mind, and with sincere respect for the different points of views in the literature and in the seminar room. I also greatly benefited from the seminars, ideas, advice, and support of Fran Hagopian and Scott Mainwaring. Fran always goes the extra mile. Even after leaving Notre Dame, she made a special effort to remain a member of my committee, and worked as hard as ever. Intellectual breadth and depth, commitment to teaching and advising, personal warmth: a perfect combination that has greatly benefited all of her lucky graduate students. Scott Mainwaring has been a key figure in my academic career since I met him at a SAAP (Sociedad Argentina de Análisis Político) conference in Mar del Plata more than twenty years ago. I had graduated from Stanford University with an MA in Political Science and another one in Latin American Studies, and, after a few years back in Argentina, was thinking about a pursuing a doctorate. Scott’s academic reputation and his leadership at Notre Dame and the Kellogg Institute were already well-established, so I was pleasantly surprised that he persistently and convincingly argued that Notre Dame would be an excellent place for me – he was right. I have been privileged to have him as a professor, advisor, coauthor, and friend. As the Director of the Kellogg Institute, as a professor of two wonderful seminars, and as a member of my dissertation committee, Scott was a central figure in my graduate school years. That he found time to teach well, advise sensibly, and publish influentially while directing a large and complex research organization is a professional feat that his students all admire. That he was at the same time demanding, respectful, supportive, generous, humble, and warm is a personal feat that we are all thankful for. I am also truly grateful to many other professors who as teachers, occasional advisors, directors of graduate studies, department chairs, or just hallway conversationalists, contributed to my education and research at Notre Dame. Benjamin Radcliff and Tim Scully taught excellent classes and provided interesting and diverse perspectives on comparative politics. Wendy Hunter and Kurt Weyland from University of Texas at Austin (UT Austin) spent a year at the Kellogg Institute, taught an amazing class on the Causes and Consequences of Market Reform, and provided those of us in that seminar with excellent academic advice. I also had the privilege of taking a class with the late Guillermo O’Donnell, Argentina’s most prestigious political scientist and a leading figure in the discipline worldwide. I was lucky to have great professors in American Politics (my second field) too: Christina Wolbrecht and David Campbell. For my methodological training I am thankful to Andy Gould, David Nickerson, Scott Maxwell (at the Department of Psychology), and Richard Williams (at the Department of Sociology). Debra Javeline gave me her expert feedback on the design of the questionnaire for the Survey of Experts on Provincial Politics.

Preface and Acknowledgments

xvii

I am thankful too for classes and academic exchanges I had with Daniel Bergan, Tim Carter, Robert Dowd, John Griffin, Alexandra Guisinger, Rodney Hero, Naunihal Singh, and Al Tillery. A special word on the Kellogg Institute and its staff. I always thought of the Institute as a second department, with its own set of guest scholars and speakers, conferences, and research funding opportunities. I made sure I made the most of the many great minds the Kellogg Institute brought to campus. Among others, I had the chance to listen to and meet with Daniel Brinks, Miguel Centeno, David Collier, Vladimir Gel’man, Clark Gibson, Kenneth Greene, Axel Hadenius, Miriam Kornblith, Margaret Levi, Steve Levitsky, Aníbal PérezLiñán, Tim Power, Donna Lee Van Cott, Lucan Way, Nina Wiesehomeier, and Deborah Yashar. The Kellogg’s staff was always helpful, efficient, and nice: at the risk of forgetting some of the many people who are or were at the Kellogg during my years there, I would like to mention Judy Bartlett, Ted Beatty, Dawn Dinovo, Therese Hanlon, Peg Hartman, July Jack, Elizabeth Rankin, Holly Rivers, Sharon Schierling, David Seymour, Christopher Welna, and Denise Wright. My gratitude goes also, and very especially, to the late Scott Van Jacob, a superb Latin American bibliographer and a great man. This book would have not been possible without the generous financial support of the University Notre Dame, its Department of Political Science, and the Kellogg Institute. The university’s Presidential Fellowship provided a generous stipend. The Graduate School awarded me a Zahm Research Travel Grant to support my field research in Argentina. The Department of Political Science provided a Grant for Research Collaboration between Faculty and students. My regular funding was complemented by the Kellogg Institute’s Supplemental Award for students from Latin America. The Kellogg also funded the exploratory stage of my fieldwork through its Seed Money Fund for Graduate Students, and then a substantial part of the main fieldwork through a Dissertation Year Fellowship. The research for this book also received the critical support of the National Science Foundation, through a Doctoral Dissertation Research Improvement Grant (Award 0719658). This funding made it possible to conduct the Survey of Experts on Provincial Politics (SEPP), the source of a very substantial part of the empirical evidence presented below. Many other colleagues helped me in one way or another, sometimes providing theoretical ideas and references, sometimes commenting on aspects of the research design, sometimes sharing their data. These include Matías Bianchi, Alejandro Bonvecchi, Valeria Brusco, Ernesto Calvo, Horacio Cao, Oscar Cetrángolo, José Antonio Cheibub, Paula Clerici, Alberto Díaz-Cayeros, Thad Dunning, Kent Eaton, Sebastián Etchemendy, Tulia Falleti, Clark Gibson, Edward Gibson, María Clelia Guiñazú, Agustina Giraudy, Lucas González, Alejandro Groppo, Matthew Ingram, Mark Jones, Carl Klarner, Frauke Kreuter, Martha Landa, Martín Lardone, Marcelo Leiras, Steve Levitsky, Lucas Llach, Germán Lodola, Nicolás Loza, Victor Mazzalay, Kelly McMann,

xviii

Preface and Acknowledgments

Osvaldo Meloni, Robert Mickey, Juan Pablo Micozzi, Kevin Morrison, Ana María Mustapic, Mario Navarro, Marcelo Nazareno, Sofanor Novillo Corvalán, Guillermo O’Donnell, Vicente Palermo, Aníbal Pérez-Liñán, Alberto Porto, Karen Remmer, Sybil Rhodes, Mario Riorda, Steven Rogers, Fernando Ruiz, Omar Samper, Tim Scully, Benjamin Smith, Catalina Smulovitz, Jorge Streb, César Tcach, Mariano Tommasi, Gloria Trocello, Laurence Whitehead, Erik Wibbels, and Javier Zelaznik. Daniel Bollada, Mercedes Llano, and Omar Samper were generous sources of contacts and support during my 2006 fieldwork in the provinces of Catamarca, Mendoza and San Luis, respectively. Marcelo Banciella, Jacqueline Behrend, Marcelo Escolar, Patricio Giusto, and Mariela Szwarczberg provided me with useful contacts in the provinces. Alejandro Cacace (San Luis), Hernán Campos (Santiago del Estero), and Lucio Conti (La Rioja) helped me obtain some hard-to-find electoral statistics in their provinces. I thank Ipsos-Mora y Araujo, and especially Santiago Lacase, Santiago Rossi, Agustina Fitzpatrick, Lucas Klobovs, and the late Manuel Mora y Araujo for making available much-needed public opinion data on presidential popularity. Much of this book, and especially the Survey of Experts on Provincial Politics, benefited from the work of great research assistants. María Marta Maroto helped me with countless aspects of the research from the project’s conception, and did so with rigor, responsibility, enthusiasm, and good judgment. Andrea Cavalli and Adrián Lucardi were critical to the success of the expert survey, and they conducted, along with María Marta, most of the interviews. The three of them traveled for long weeks to cover each of Argentina’s twenty-three provinces, and did a superb job of contacting and interviewing dozens of experts. They also coded the questionnaires, always working with great professionalism and sagacity. In Buenos Aires, María Eugenia Wolcoff assisted me in coordinating the fieldwork and conducted several interviews, doing so with dedication, congeniality, and youthful enthusiasm. Julieta Altieri, Tomás Bieda, Victoria Romano Florit, Jimena Sánchez, and Geraldine Sznek provided excellent research assistance to collect secondary data on the eight countries included in the comparative Chapter 7. Nicole Dipp and Federico Méndez Casanave assisted me in collecting and organizing bibliographic and journalistic evidence about Argentina’s provincial regimes. Tomás Listrani carefully and skillfully designed the three maps that appear in the book. I am very thankful to all of them. I owe much to the 155 experts interviewed for the Survey of Experts on Provincial Politics. Their names (except for a few who preferred to maintain their anonymity) can be found in the online appendix.1 Much of the knowledge generated by the survey would have been impossible without their own knowledge and willingness to share it. I conducted much of the research for this book while I was working for universities in Argentina between 2008 and 2018. During 2008, I taught at 1

 At www.utdt.edu/profesores/cgervasoni.

Preface and Acknowledgments

xix

Universidad Católica Argentina, Universidad del CEMA, Universidad de San Andrés, and Universidad Torcuato Di Tella, and in all of them I found good colleagues to discuss my ideas with. In 2009, I started a full-time position at the Department of Political Science and International Studies at Universidad Torcuato Di Tella, where I had also taught for several years before I left for my Ph.D. at Notre Dame. I am grateful for the intellectual stimulation and pluralism that I have always found at the department and the university. I am very proud to belong to them. My wonderful family has had a huge role in this book. My mother Kety and my father Carlos (who passed away in 2013) are, in many ways, the distal causes of it – not only because of their support and love throughout my life, but because of their moral education and example. I was lucky to be able to spend much time with Carlos before he became ill with cancer, and during the tough months that ensued. In the sadness and pain of the last days, he remained calm, sober, and selfless. He left teaching those around him a lesson in strength, endurance, and dignity. I admired him and loved him very much. My wife Sybil, who happens to be an accomplished political scientist, supported me in every way and gave me advice about countless aspects of the book. She has been, throughout our many years together, a loving partner and a kind heart. Federico, our son and our masterwork, grew together with the ideas I present in this book. Inevitably he has been, like Arthur Stinchcombe’s wife, “alternately a hindrance and a help,” but always a great source of pride, satisfaction, and joy – and the lucky target of tons of care and love. As I was putting the final touches on these pages, he repeatedly came into my home office to ask, “Hey Daddy, how is the book going?” I was lucky to have Sara Doskow as my editor at Cambridge University Press. As the original manuscript went through the review process and eventually turned into this book, Sara was consistently helpful, diligent, and affable. Along with the quick and useful answers provided by her editorial assistants Claudia Bona-Cohen and Danielle Menz, I felt I could always count on professional and timely answers to any inquiries. The book also benefited from the professional work of Cambridge’s Claire Sissen, its content manager, and Paris West, its marketing associate. Three anonymous reviewers provided thorough, sensible, and insightful suggestions that helped improve the original manuscript. Geetha Williams and Ezhilmaran Sugumaran, from MPS Limited, managed the production of the book. My gratitude goes to all of them. Some of the research I conducted for this book has been published in the form of journal articles and book chapters. In the early stages I wrote some pieces in Spanish, which were published in the Argentine journals Colección, Boletín de Política Comparada, and Aportes para el Estado y la Administración Gubernamental. More recently I published some of my descriptive and explanatory findings in World Politics (Gervasoni 2010a), the Journal of Politics in Latin America (Gervasoni 2010b), and in two book chapters (Gervasoni 2016a, 2016b). Portions of the book draw on these pieces.

Abbreviations

ACLP CSDI CM NK PJ RA SDI SEPP UCR V-Dem

Alvarez, Cheibub, Limongi and Przeworski Comparative Subnational Democracy Index Carlos Menem Néstor Kirchner Partido Justicialista (aka Partido Peronista or Peronism) Raúl Alfonsín Subnational Democracy Index Survey of Experts on Provincial Politics Unión Cívica Radical (aka Partido Radical or Radicalism) Varieties of Democracy Project

Provinces PBA Buenos Aires CF Capital Federal (or Ciudad Autónoma de Buenos Aires, or CABA) CAT Catamarca CHA Chaco CHU Chubut CBA Córdoba CTS Corrientes ER Entre Ríos FSA Formosa JUJ Jujuy LP La Pampa xxi

xxii

List of Abbreviations

LR La Rioja MZA Mendoza MIS Misiones NEU Neuquén RN Río Negro STA Salta SJ San Juan SL San Luis SC Santa Cruz SF Santa Fe SE Santiago del Estero TF Tierra del Fuego, Antártida e Islas del Atlántico Sur TUC Tucumán

Introduction*

A full description of the opportunities available for participation and contestation within a country surely requires one to say something about the opportunities available within subnational units. – Robert Dahl (1971, 12)

The Peronist party has ruled the Argentine province of Formosa since the re-­democratization of the country in 1983, winning nine consecutive gubernatorial elections. The current governor, Gildo Insfrán, was first elected in 1995 with 59 percent of the vote, and then reelected five times in a row with between 72 percent and 76 percent of the vote. Formosa has a very weak system of checks and balances: neither the provincial legislature (dominated by a large Peronist majority) nor the provincial judiciary control the governor.1 How was this level of hegemony achieved in a province with some of the country’s worst social development indicators? Some readers may think poverty breeds this kind of undefeatable local bosses, but one can find many regions in Argentina and other countries that are similarly underdeveloped and yet regularly produce competitive elections and rotation of parties in office. Other readers may suspect massive levels of patronage and clientelism, and they would be right: 54 percent of all formal workers in Formosa are on the provincial payroll (and an additional 9 percent are municipal employees).2 Exaggerating only slightly, everybody is a provincial public employee or lives with someone who is. Therefore, most Formosans depends on the provincial budget for their economic survival. Can democracy function when the ruled are employees of the rulers, rather than the rulers being employees of the ruled? *

  Portions of this chapter were published previously in Gervasoni (2010a).

1

 See Chapter 3 and in particular Figure 3.8.  Pozzo (2017).

2

1

2

Introduction

A second intriguing question: how can such a poor province pay for so many employees? Formosa’s own tax revenues are tiny, accounting for barely 1 percent of its gross geographic product and just 5 percent of its public expenditures.3 The other 95 percent comes from federal transfers sent by the national Treasury in Buenos Aires. The answer to the question, then, is that Formosa’s employees are paid with taxes raised by the federal government in the rest of the provinces (especially the largest ones in population) and redistributed in favor of demographically small provinces like Formosa. I will argue that there is a link between Insfrán’s capacity to spend large amounts of money on public employees (and other things) without really taxing local constituents, and his long, hegemonic rule. Plentiful federal transfers to provinces like Formosa are subsidies, or unearned income, which, much like resource rents in oil-­rich countries, give rulers the capacity to make most individuals, groups, and companies dependent on the state’s budget without the state budget being dependent on them. It is not poverty, but the capacity to spend much money without taxing local actors that allows rulers like Insfrán to obtain electoral landslides unheard of in truly democratic polities (including Argentine provinces that do not enjoy high levels of federal subsidies). What I will call “fiscal federalism rents” are detrimental to subnational democracy in the same way that, according to many academic studies, oil rents are detrimental to national democracy. Contextual details vary, but the causal logic at work and the mechanisms connecting high rents with low democracy are essentially the same. Why does subnational democracy (or the lack thereof) matter? In federations – and increasingly in formally unitary countries with high levels of decentralization – lower levels of government wield much power. Many of the laws, and most of the officials, judges, policemen, doctors, and teachers who make decisions and supply public services, belong to (first-­level) subnational units such as US states, German länder, and Argentine provinces. Even if the national regime is democratic, autonomous regional governments command considerable fiscal, human, and bureaucratic resources which allow them to curtail political rights in many and important ways. Consider the case of the Southern US states, which were able to sustain racial segregation and one-­party regimes for many decades in spite of being part of one of the oldest and most-­consolidated national democracies in the world. Less democratic subnational regimes show their ugliest face when they resort to visible coercion. I will argue, however, that high profile acts of repression are rare in such regions when they are embedded in a national democracy. Subnational rulers with authoritarian inclinations understand that banning a critical newspaper, crushing an opposition protest, or jailing a troublesome judge are risky moves in national democratic contexts. Instead, they will try to conceal their authoritarian ways by, for example, withdrawing publicity from the critical newspaper, signaling that participation in a protest will lead

3

 See Table 5.1.

Introduction

3

to losing a public job, or blackmailing the judge after spying on her or his private life. These “soft” forms of repression significantly undermine democratic rights and freedoms, as media outlets “choose” not to publish facts or opinions unfavorable to the provincial administration, citizens fear turning up at an opposition rally, and judges default on their obligation to control the governor’s actions. Although the specific tactics provincial government use are difficult to observe, their consequences are not: Chapter 3 documents, among other things, the prevalence in several Argentine provinces of grossly biased media systems, of low levels of freedom of expression for public employees, and of very weak checks and balances. These practices, paraphrasing a scholar of the province of San Luis, manufacture “serf citizens” (Trocello 2008; author’s translation). Many elements central to mainstream definitions of democracy – freedom, accountability, limitations on the power of the executive, and a level electoral playing field – are critically weakened by them. These failings of democracy at the subnational level would be grave enough if they just affected the citizens of a few provinces. The experience of Argentina, however, shows that a country’s least democratic regions often have a disproportionate impact on its national politics. In particular, the governors of those provinces have been remarkably successful at capturing the presidency in Argentina. Carlos Menem (CM) and Néstor Kirchner (NK) won presidential elections in 1989 and 2003, respectively, when they were serving as governors of La Rioja and Santa Cruz, both demographically tiny provinces with some of the lowest levels of democracy in the country. Many of the hegemonic tendencies in their presidencies (and those of Kirchner’s wife and successor, Cristina Fernández) can plausibly be attributed to their political socialization in low-­ democracy environments. This point was elaborated in the Preface as my frustration with the authoritarian tendencies of presidents Menem and Kirchner was an important motivation for undertaking the research project that led to this book. I quote from there: “Menem, for example, quickly packed the Supreme Court with a group of loyal and often unqualified cronies that made a joke of judicial checks and balances. Kirchner, among other things, targeted one of the most basic pillars of democracy, the critical media. No wonder: there are hardly any independent judges in La Rioja and hardly any independent media outlets in Santa Cruz. The first instinct of these leaders was to reproduce at the national level the hegemonic realities they had become used to in their provinces.” George Wallace, the proudly segregationist governor of Alabama, ran four times for the presidency. One wonders how civil rights, and ultimately democracy, would have fared in the United States had he been successful. Beyond becoming presidents, the rulers of the least-­democratic regions affect the national regime in other ways, for example through their influence on legislators elected in their territories, who often represent the governors’ interests in the national legislature. Jim Crow era Southern Democrats in the US Congress, as well as Peronist deputies and senators from Formosa, La Rioja, and Santa Cruz have often helped undermine national democracy. For reasons ranging from the seniority system in the US Congress to gross malapportionment in

4

Introduction

Argentina, these legislators often command much more power than the relative demographic magnitude of their regions would suggest.

I.1  Getting on the Agenda: Comparative Politics and Subnational Regimes Scholars of political regimes have long noticed that the extent to which citizens of democracies enjoy political rights and freedoms varies widely, not only across social cleavages such as class and ethnicity, but also across subnational/ regional4 boundaries. The United States during the “Solid South” years, when a robust national democracy coexisted with a group of racially e­ xclusionary, single-­ party subnational regimes (Key 1949; Gibson 2012; Mickey 2015), provided an early and stark example. In his classic Polyarchy, Robert Dahl noted, “even within a country, subnational units often vary in the opportunities they provide for contestation and participation” (1971, 14), adding that not dealing with this issue was a “grave omission” (p. 12) of his book. Four decades into the “third wave,” it is even clearer that many national democracies include some very imperfectly democratic subnational regimes. The problem appears to be especially acute in large and diverse federations in the developing world. Countries such as Argentina, Brazil, India, Mexico, and Russia show remarkable heterogeneity in the degree to which their regions are democratic. Nevertheless, as the example of the US South shows (and the results of a comparative analysis confirm in Chapter 7), established democracies in developed nations are not necessarily free from the problem. Political science is just starting to remediate Dahl’s omission. Much of the burgeoning literature on federalism, decentralization, malapportionment, and other dimensions of subnational politics does mention the existence of authoritarian practices at the regional level, but usually just as a passing remark. Despite massive research on national authoritarianism and democracy in the last six decades, there are only a few such studies at the subnational level. Standard and widely used indices of democracy – for example that produced by the Polity IV project – do not consider any subnational information to code countries. The ambitious Varieties of Democracy project – a research initiative to improve the measurement of national regimes – does, for the first time to my knowledge, include a regional dimension (Coppedge, Gerring et al. 2011; Coppedge et al. 2018a, b).

 I refer to autonomous polities within sovereign countries (and to their regimes) using three terms at different levels of abstraction. The most general one is “subnational,” as in “subnational unit” or “subnational regime.” When referring to first-­level subnational units, such as Argentine provinces, German länder, Russian republics or oblasts, or US states I use “region” and “regional regimes.” Finally, when the text refers specifically to Argentina I use “province” and “provincial regime.”

4

Introduction

5

The topic timidly entered the research agenda of comparative politics in the 1990s, as the functioning of the new third-­wave democracies significantly expanded the number of formally democratic subnational regimes. This new source of evidence eventually made it quite clear that within-­country regime variance was substantial. Several scholars highlighted the existence of subnational “authoritarian enclaves” (Fox 1994, 106) or the persistence of authoritarian “traditional politics” (Hagopian 1996) at the regional level, even after national transitions to democracy. Toward the end of the decade, the topic figured with some prominence in the works of influential scholars of democracy. They alerted that “[e]nclaves of exclusion and repression exist throughout Latin America and the successor states of the Soviet Union” (Diamond 1999, 133) and wondered “how one conceptualizes a polyarchical regime that may contain regional regimes that are not at all polyarchical” (O’Donnell 1999b, 315). The introduction to the edited volume Subnational Politics and Democratization in Mexico called attention to the “survival and even strengthening of subnational authoritarian enclaves in states like Puebla, Tabasco, Guerrero, Chiapas, Oaxaca, Campeche, and Yucatán” and, more generally, to the “uneven, patchwork character of democratization” in Mexico (Cornelius 1999, 3–4). Mexico’s decisive move toward national democratization in 2000 was considered an “unfinished transition” because of the persistence of these enclaves (Lawson 2000). Not coincidentally, an early and influential piece on the methodology of subnational comparisons used as one of its main illustrations the persistence of “illiberal peripheries” in new national democracies (Snyder 2001, 101–2). In Argentina, journalists preceded academics in throwing light on “problematic” provincial regimes. Soon after the country returned to democratic rule in 1983, a few investigative journalists started writing about provinces such as Catamarca, San Luis, and Santiago del Estero where elected Peronist governors seemed to exercise power well beyond the limits of democratic principles. Local scandals that made the national news (such as the 1990 “María Soledad” case in Catamarca5) and colorful provincial bosses (e.g., governors Adolfo Rodríguez Saá of San Luis and Carlos Juárez of Santiago del Estero) aroused the interest of journalists. The resulting stories and books focused on many maladies, such as personalism, nepotism, clientelism, and corruption, but authoritarianism (not necessarily under that name) was a paramount topic of these works. A sample of excerpts from the titles and subtitles of these 5

 María Soledad Morales, 17, died in obscure circumstances, apparently during a party organized by people linked to the provincial regime. Eventually Guillermo Luque, the son of a Catamarcan Peronist federal deputy, was found guilty of her rape and murder, and sentenced to twenty-­one years in prison (he was released in 2010, after fourteen years, for good behavior). The case became a provincial and national scandal, especially after clear signs that the provincial government headed by Ramón Saadi was trying to cover up friends. Massive and repeated demonstrations kept the scandal alive for months. In 1991, President Menem used the constitutional power of “federal intervention” to remove Saadi from the governorship.

6

Introduction

books – even untranslated – is instructive: Historia de un Feudo (Zicolillo and Montenegro 1991, on Catamarca), Cuando el Tirano cae su Poder Termina (Morandini 1991, on Catamarca), El Último Feudo (Wiñazki 1995, on San Luis), Contrademocracia Argentina (Bazla 2002, on San Luis), Crónicas del Fascismo Mágico (Wiñazki 2002, on San Luis), Miseria, Terror y Desmesura (Carreras 2004, on Santiago del Estero), and Terror, Corrupción y Caudillos (Dandan, Heguy, and Rodríguez 2004, on Santiago del Estero). Even allowing for some measure of attention-­grabbing exaggeration, these titles are telling: one does not use words such as “feudal,” “tyrant,” “fascism,” and “terror” to describe regimes that are reasonably democratic. These and other journalistic works helped put provincial regimes on the agenda of scholars of Argentine politics and provided them with an initial and valuable, if nonsystematic, body of empirical evidence to start exploring the subject. Later academic works confirmed that these journalists were on to something. Further research reveals that Argentina has one of the world’s highest levels of regional unevenness in the fairness of its subnational elections and in the respect for civil liberties (McMann et al. 2016). This should not be surprising, as the country has medium to high values in most of the variables that, according to the cited study, tend to predict regime heterogeneity across regions: terrain ruggedness, population size, and, especially, economic inequality among regions. Moreover, because Argentina is a federal nation with very powerful first-­level subnational units (Hooghe et  al. 2016), what journalists were observing was very real: a few provincial governments that were (de facto) much less democratic than others and (de facto and de jure) very autonomous to make their own decisions over a wide range of policy areas. One of the main data sources for this book is a survey of experts on the politics of each Argentine province. An item toward the end of the questionnaire supplied the experts with a definition of democracy and asked them to rate the reference period (2003–7) in their provinces (plus the Raúl Alfonsín, CM, and NK national administrations) on a scale ranging from “very democratic” to “not democratic at all.”6 This item is not the most rigorous measure of subnational democracy because it leaves much to the different criteria that different experts may use,7 but it is useful as a simple, “quick and dirty” empirical  The question read as follows: “For the next questions I need to define democracy as ‘a political regime in which: (1) the executive and legislative branches are elected in free and fair elections with universal adult franchise, (2) there are effective checks and balances among the executive, legislative and judicial branches, and (3) basic constitutional rights such as freedom of speech are respected.’ I am going to mention several provincial and national governments, and I would like you to tell me, using this definition, whether each of them was very democratic, quite democratic, somewhat democratic, not very democratic or not democratic at all.” The wording in Spanish can be consulted in the online appendix at www.utdt.edu/profesores/cgervasoni. 7  For example, although a definition is given, experts may still bring into their assessments their own ideas about what democracy is. Even those who strictly follow the definition may interpret the three elements included in it differently, and weight them in different ways. 6

Introduction

7

1

0.6

0.4

0.2

0

SC SE SL STA JUJ LR MIS FSA CM NK CHA NEU CHU TF TUC CAT LP RN CBA SJ PBA ER RA CTS SF MZA CF

Overall democracy (2003–7)

0.8

Figure I.1.  Overall evaluations of subnational democracy in Argentina (2003–7)

exploration of our subject matter. Figure I.1 presents the provincial experts’ average scores on this overall indicator of democracy (0 = minimum level of democracy; 1 = maximum level of democracy). The main point to notice is that provincial democracy varies significantly: the city of Buenos Aires (CF), Mendoza, Santa Fe, Corrientes, Entre Ríos, Buenos Aires, San Juan, and Córdoba are deemed reasonably democratic by experts while, at the other end, Santa Cruz, Santiago del Estero, San Luis, Salta, Jujuy, La Rioja, and Misiones are seen as much less democratic. The national-­ level ratings (white columns) provide a useful (and interesting in itself) point of reference for comparison. As Figure I.1 shows the 1983–9 Raúl Alfonsín (RA) administration was considered democratic by the 155 experts (although less so than the most democratic provinces). The CM (1989–99) and NK (2003–7) administrations were evaluated as considerably less democratic, and in fact below most provinces. Again, these are not the best data to describe subnational democracy in each province (e.g., Corrientes and Formosa do significantly worse with the alternative measures used in the rest of this book), but the figure as a whole shows two real patterns: (1) that provinces vary widely, from very democratic to rather undemocratic; and (2) that they can be, as expected, less democratic than the national regime but, surprisingly, also more democratic. Argentina is not atypical. According to V-­Dem data, as of 2012, 68 percent of all countries in the world had elected regional executives and/or assemblies, but unevenness within them in terms of the fairness of the elections and the

Introduction

8 2

Subnational elections unevenness

Somalia

Russia

Nigeria

Iraq

Argentina

Ethiopia

1

Mexico

Pakistan

Brazil

Malaysia Sudan

Venezuela

India

0

United States of America

–1

–2

Australia Canada Germany Switzerland Austria

Adj. R2 = 0.53

–2

–1

0

1

2

3

Subnational elections free and fair

Figure I.2.  Subnational elections in nineteen large federations (2014) Source: Author’s elaboration on varieties of democracy data (Coppedge et al. 2018a)

respect for civil liberties was often high (McMann 2018). Figure I.2 presents indicators of regional elections’ freeness and fairness (X-­axis) and of their territorial unevenness (Y-­axis) for nineteen large federal countries. There is significant variance in both axes. Some countries have subnational elections that are largely free and fair (e.g., Germany and the United States) some leave much to be desired (e.g., Somalia) and some obtain middling levels (such as Argentina, India, Mexico, and Nigeria). Likewise, in some nations elections are of similar quality across subnational units, while in others there is significant variance. The two variables are quite strongly correlated (adjusted R2 = 0.53), indicating that those countries that have freer and more fair regional elections tend to also be more even territorially. Nations such as Brazil, Mexico, Nigeria, the United States, and especially Argentina, are above the regression line, indicating that the quality of subnational elections there is more uneven than expected on the grounds of their overall subnational elections quality. Not surprisingly, these are among the countries most studied by the literature on subnational democracy: differences within countries are especially visible and especially puzzling, given that many potential explanatory factors vary little (or not at all) across the regions of a given nation. These differences among the regions of a given country call for an explanation. An important advantage enjoyed by scholarship on subnational regimes is that it can “stand on the shoulders of national giants,” so to speak: political science has produced, since Lipset’s (1959) seminal article, a huge, theoretically rich, and methodologically sophisticated literature on measuring and explaining

Introduction

9

democracy and authoritarianism at the national level.8 The recent and small literature on subnational democracy, then, has a solid foundation on which to build, as well as great potential to contribute better measurement tools, innovative theories, and critical evidence on the extent of, and reasons for, regime differences. Expanding scholarship on democracy from the national to the subnational level greatly increases the size and diversity of the relevant units of analysis. The thousands of regions, territories, counties, and municipalities around the world are an exceptional source for new ideas and new evidence. Over the past two decades, scholars have produced several studies of subnational democracy, or of related concepts such as democratic competitiveness or human rights respect. Qualitative case studies have been conducted for Argentina (Chavez 2003, 2004; Gibson 2005, 2012; Trocello 2008; Behrend 2011; Giraudy 2015), Brazil (Hagopian 1996; Durazo Herrmann 2014; Borges 2016; Souza 2016), India (Heller 2000; Tudor and Ziegfeld 2016), Kyrgyzstan (McMann 2006), Mexico (Cornelius, Eisenstadt, and Hindley 1999; Snyder 1999; Gibson 2005, 2012; Durazo Herrmann 2010; Giraudy 2015), the Philippines (Sidel 2014), Russia (Gel’man et al. 2003; Petrov 2004; McMann 2006), South Africa (Munro 2001), and the United States (Gibson 2012; Mickey 2015; Gibson and King 2016). There have also been several quantitative studies aimed at describing and/or explaining subnational regimes within a single country, for Argentina (Gervasoni 2010a, 2010b, 2016b; Giraudy 2010, 2015), Brazil (Borges 2016), Mexico (Hernández Valdez 2000; Beer and Mitchell 2004; Giraudy 2010, 2015; Somuano Ventura and Ortega Ortiz 2011; Gervasoni 2016b; Loza and Méndez 2016), India (Beer and Mitchell 2006; Lankina and Getachew 2012), Russia (McMann and Petrov 2000; Lankina and Getachew 2006; Saikkonen 2016a, 2016b), and the United States (Hill 1994; Goldberg, Wibbels, and Mvukiyehe 2008). There are also quantitative studies conducted at the municipal level (Gel’man and Lankina 2008; Benton 2012). More recently and taking advantage of the new Varieties of Democracy dataset, scholars have begun to study subnational regimes across the world (McMann et al. 2016), transcending the previous focus on one or a few countries. These are significant contributions, but they pale in comparison to the exemplary literature on national regimes, characterized by high and increasing levels of theoretical sophistication and methodological rigor. Building both on the national and subnational literatures, this book attempts to advance our knowledge by: (a) making an in-­depth descriptive and explanatory contribution for the specific domain of Argentina’s twenty-­four first-­level subnational units; and (b) providing a comparative perspective with several other countries, hoping (and claiming) that the book’s methodological and theoretical innovations are applicable to many other nations and many

 Too large to cover adequately here; for an excellent summary see Coppedge (2012).

8

10

Introduction

other subnational units. In the next two sections, I provide a summary of the book’s original descriptive and explanatory contributions, which are developed at length in the following chapters.

I.2  Description: Objective and Subjective Operationalizations of Subnational Democracy There exist many and diverse indices of democracy (or regime type) at the national level, some of which are longstanding and cover most countries of the world over extended periods of time (such as the widely ­used Polity score). That is not the case for the subnational level. Therefore, a substantial part of this book is dedicated to: (a) developing several measures of subnational democracy; (b) applying them to all the provinces of Argentina to provide a comprehensive description of their political regimes; and (c) applying one of these measures comparatively to eight countries: Argentina, Australia, Canada, Germany, India, Mexico, the United States, and Uruguay. This descriptive effort is valuable in itself, and constitutes an indispensable foundation for causal inference. For Argentina I developed two alternative operationalizations of subnational democracy, one objective (which results in the Subnational Democracy Index, based on electoral and institutional indicators) and one subjective (which results in several indices derived from a survey of local experts on provincial politics), and applied them to all the provinces for the periods 1983–2015 and 2003–7, respectively. The resulting indices provide a rich and detailed anatomy of provincial regimes in contemporary Argentina, covering many aspects of democracy, from the core electoral competition component to the status of press freedom, the effectiveness of checks and balances, and the prevalence of human rights violations. For the rest of the countries analyzed in this book, I developed a generalized version of the objective index – the Comparative Subnational Democracy Index (CSDI) – that can be applied to nations with different institutional characteristics than Argentina. This index is relatively narrow in the sense that it taps only the “contestation” dimension of democracy, but has the advantage of permitting both cross-­unit comparisons among different countries and within-­unit comparisons over time. Two key conclusions emerge from these data: (1) regions do vary considerably in the extent to which they are democratic, and do so along many dimensions of political regimes; but (2) they typically range from democratic to hybrid, hardly ever reaching the authoritarian pole in countries where the national regime is reasonably democratic. Paralleling national-­level findings, regimes that combine elements of democracy and authoritarianism appear to be, in certain contexts, more viable, and therefore more prevalent than outright authoritarianisms (Zakaria 1997; Levitsky and Way 2002, 2010; Ottaway 2003; Schedler 2006).

Introduction

11

The subjective operationalization strategy was carried out through the Survey of Experts on Provincial Politics (SEPP), to my knowledge the third attempt (after those of Kelly McMann and Nikolai Petrov in Russia and Kyrgyzstan; McMann and Petrov 2000; McMann 2006) to measure subnational democracy in a country via an expert survey.9 The SEPP has a number of methodological strengths (such as its detailed questionnaire covering many dimensions of democracy, its consulting several experts per province, its reliance on province-­specific, province-­based experts) that together make it, I believe, the most significant contribution to the measurement of subnational democracy to date. Such expert-­survey methodology can easily and fruitfully be applied to other regional regimes around the world (the SEEP has already being replicated in Mexico10), as well as to national regimes. A descriptive strength of this book is the measurement of the same subnational regimes through two independent operational strategies based on very different informational sources (objective indicators and expert judgments). Descriptive inference is inevitably uncertain, but this uncertainty is reduced when empirical findings are confirmed by independent methodologies. Some Argentine provinces are very imperfectly democratic, and some quite democratic, under either measurement strategy. At least for these districts, there should be few doubts about the nature of their regimes. Causal inference also benefits from these alternative measurements strategies: hypotheses that hold well under tests using one or the other are more likely true than those that are sensible to the way the dependent variable is measured.

I.3  Explanation: A Rentier Theory of Subnational Democracy As for explanation, I present and test a theoretical account of subnational regimes that centers on a rentier understanding of certain strains of fiscal federalism. In its briefest form the argument (developed in detail in Section 5.1) is that certain varieties of fiscal federalism (such as Argentina’s) imply that some regions receive huge federal subsidies, which, working as functional equivalents of oil rents, give rise to rentier subnational states. The rulers of these regions, rich in fiscal resources, autonomous from taxpayers, and in command of an economically dominant state, are especially likely to succeed in undermining democratic limitations on their power. Simply put, I argue that “fiscal federalism rents” hinder subnational democracy in the same way that oil rents hinder national democracy (Ross 2001).

 Petrov and Titkov (2013) also produced an index of democracy for Russia’s regions based on expert opinions. 10  See EEPEMEX, or Encuesta a Expertos en Política Estatal en México (Loza and Méndez 2016) https://podesualflacso.wordpress.com/2013/12/18/eepemex/. 9

12

Introduction

Rents are “income accruing to a state from an external source on a regular basis, that far exceeds the cost incurred to obtain it, and that involves the production effort of a very small proportion of the state’s population.”11 Why might they be detrimental to democracy? The typical rentier state (say a small Persian Gulf nation) obtains rents by exporting millions of barrels of oil at prices much higher than their production cost. Rulers of rentier states can sustain high levels of public spending with low levels of taxation. Citizens, groups and businesses become dependent on the plentiful government’s treasury, while the treasury hardly depends on them. The large literature on the political resource curse (reviewed in Section 4.2.3) posits that, under these conditions, democracy is unlikely to emerge or survive. A critical question is the following: would anything change if the source of rents was something else instead of oil or other similar natural resources? The answer this book provides is a resounding “no, nothing important would change.” What is critical about rents is not their origin (notice the definition above does not mention a specific source), but the fact that they give the state and the elites in charge of it an enormous political benefit: the ability to reap the rewards of spending without bearing the costs of taxing. Governments in this advantageous fiscal position command so much power that they are unlikely to face (or yield to) democratic pressures from constituents. Even if democratic institutions are in place, the level of incumbency advantage provided by a large and externally funded budget is hardly compatible with a reasonably level electoral playing field. Democracy is more formal than real if the incumbent is essentially undefeatable. Many regions around the world (including several Argentine provinces) are subnational rentier states because they receive plentiful “fiscal federalism rents.” That is, the governments of these regions are heavily subsidized by the national government (and, indirectly, by the taxpayers of other regions). They may not enjoy resource rents (some do), but they certainly receive substantial “income . . . from an external source on a regular basis, that far exceeds the cost incurred to obtain it, and that involves the production effort of a very small proportion of the population.” This is certainly the case of Argentine provinces such as Catamarca, Formosa, La Rioja, Santa Cruz, and Santiago del Estero, notorious for receiving federal transfers far beyond what their local tax bases could produce. The details of the intergovernmental revenue-­ sharing rules that lead to this kind of rentierism are spelled out in the first two sections of Chapter 5, but I anticipate here that the facilitating conditions are large vertical fiscal imbalances combined with highly disproportional criteria for distributing federal funds among subnational units. This type of fiscal federalism gives rise to rentierism because all regional governments depend heavily on federal transfers for funding their (typically large) budgets, and because the subset benefited by

 I develop and justify this definition in Section 4.5.

11

Introduction

13

the distribution criteria receive transfers that far exceed their contribution to the national tax pool. The governments of these provinces – like those of oil-­ rich emirates – live in a political heaven: they can sustain high levels of fiscal spending without significantly taxing their constituents. The hefty and externally financed provincial budget becomes the center of the local economy. Not only are most economically active residents public employees, but provincial companies, media outlets, and nongovernmental organizations find themselves depending heavily on government contracts and subsidies. These actors are likely to refrain from behaviors that may upset the government that issues their paychecks, such as criticizing it or supporting opposition parties. Democracy in unlikely to thrive when the incentives are so heavily biased in favor of the incumbent. Thus, as described in the first lines of this chapter, Peronist Governor Gildo Insfrán, of the highly rentier province of Formosa, has little trouble winning reelection after reelection with over 70 percent of the votes. In summary, the theory expects that regions whose governments control very large budgets mostly financed through plentiful federal transfers will develop “rentier state” features that, in turn, will undermine democratic rule. This idea, derived inductively from the Argentine provinces, fits very well with four important literatures in political science: (1) fiscal theories that emphasize the political consequences of the way states are financed (Bates and Lien 1985; Levi 1988; Karl 1997; Bräutigam, Fjeldstad, and Moore 2008; Morrison 2009, 2015); (2) the rentier-­state approach, which postulates a negative causal impact of resource rents on democracy (Mahdavy 1970; Beblawi and Luciani, 1987; Ross 2001); (3) the liberal tradition that posits a normative and empirical tension between extreme economic statism and democracy (Hayek 1944; Dahl 1971; Sartori 1987; Diamond, Linz, and Lipset 1995); and (4) the literature on the political consequences of different varieties of fiscal federalism (for the case of Argentina see Gibson and Calvo 2000; Remmer and Gélineau 2003; Calvo and Murillo 2004; Wibbels 2005; Remmer 2007; Gervasoni 2013; for other countries see, for example, Rodden 2006; Brollo et al. 2013; Morrison 2015, Chapter 5). In agreement with these four literatures, the rentier theory of subnational democracy I advance: (1) postulates that the way provincial governments are financed is critical to understanding their politics; (2) posits that “fiscal federalism rents” are detrimental to subnational democracy; (3) predicts more authoritarian politics where the provincial state dominates economic activity; and (4) highlights that the rentier syndrome that weakens democracy in several provinces is a result of Argentina’s intergovernmental revenue-­sharing rules (or fiscal federalism institutions). Notoriously absent in this theoretical account is modernization theory, a central paradigm in democracy scholarship. My rentier approach says nothing about the causal effect of economic development and related social transformations. These factors appear in this book as controls. If the rentier theory of subnational democracy is broadly correct, we should expect to find lower

14

Introduction

levels of democracy in the provinces most “benefited” by fiscal federalism rents after controlling for development. A national-­ level illustration is in order: Qatar is among the world’s top five countries in terms of GDP per capita, well above most developed countries, and has a human development index comparable to that of Greece and Slovakia. That in spite of these “preconditions” it never joined the third wave of democratization, many believe, has something to do with its being the world’s largest per capita producer of oil. Argentina contains some subnational Qatars. Santa Cruz, for example (the province that Néstor Kirchner ruled for twelve years before becoming president in 2003), is among the richest in the country, but it is also rentier (because of plentiful fiscal federalism and oil rents) and, unsurprisingly, hardly democratic. Rentierism might be strong enough to trump the alleged positive effects of development on democracy or, alternatively, likely to produce a statist development pattern whose causal effect on democracy is the opposite of that expected by the classic capitalistic development road envisioned by modernization theory.

I.4  Democracy, Federalism, and Subnational Politics in Argentina Argentina is the eight largest country in the world by land area, the thirty-­first largest in population, and the twenty-­first largest in economic size (GDP). It became independent from Spain in the early years of the nineteenth century, and after several decades of political turmoil and civil war, it adopted a presidential and federal Constitution that, with several partial reforms (and interruptions due to military governments), has lasted until today. During the last decades of the nineteenth century and the first years of the twentieth century, the country stabilized politically (as a sort of oligarchic liberal republic), prospered economically, and attracted millions of mostly Spanish and Italian immigrants. By 1912 the ruling elites yielded to pressures for a more democratic franchise, which in 1916 led to the first competitive presidential election with secret and universal (male) suffrage, and to the first electoral rotation in the presidency: the Unión Cívica Radical (UCR), headed by Hipólito Yrigoyen, defeated the long ruling Partido Autonomista Nacional. This democratic progress, however, was interrupted by a military coup in 1930 (which ousted Yrigoyen during his second presidency). Over the next five decades, the country underwent a period of acute regime instability, marked by successive democratic transitions and military coups (one of the latter was the cradle of Peronism, the dominant political force in Argentina since the 1940s). In fact, during the 1950–90 period studied by Przeworski et al. (2000), Argentina was the country with the most transitions to and from democracy in the world: eight in total. After the first 1930 coup the country suffered five more military coups (in 1943, 1955, 1962, 1966, and 1976) and six transitions to electoral regimes (in 1932, 1946, 1958, 1963, 1973, and 1983). Using three regime categories  – democracy, semi-­democracy, and authoritarianism – Mainwaring and Pérez-­Liñán (2013)

Introduction

15

find that between 1900 and 2010 Argentina experienced sixteen regime transitions.12 Civilian regimes were not necessarily synonymous with full democracy, as some presidents were “elected” in rigged contests (during the 1932–43 period) or in elections in which a major party (Peronism) was banned (in 1958 and 1963). Moreover, one transition to authoritarianism occurred during the stint of a democratically elected government: Colonel Juan Perón – a central figure of the 1943–6 military regime – won free elections in 1946, but went on to establish, first, a semi-­democratic regime, and after 1951, an authoritarian one (Mainwaring and Pérez-­Liñán 2013). Argentina’s regime turbulence ended in 1983, in the midst of the world’s third wave of democratization (Huntington 1991). The military regime inaugurated in 1976 fell into total discredit in the early 1980s, under the combined weights of an economic crisis, mounting evidence of atrocious human rights violations, and military defeat in the 1982 Malvinas/Falklands War. Democracy was reestablished at the end of 1983, when Raúl Alfonsín (the UCR presidential candidate) defeated Peronist Ítalo Lúder in free and fair elections. Although the country has suffered very significant political and economic crises since then, the national political regime has remained essentially democratic. The military ceased to be a real threat for democracy, while “threats from within”  – attempts by Peronist presidents Carlos Menem in the 1990s and Néstor and Cristina Kirchner in the 2000s to establish hegemonic electoral regimes – failed. Electoral rotations in the national executive took place peacefully, in 1989, 1999, and 2015, and even the resignation of President De la Rúa (from the UCR) in the midst of a dramatic economic crisis in 2001 was processed institutionally, with the election by Congress of a new interim president, Peronist Eduardo Duhalde. Argentina’s post-­1983 democratic endurance seems to have been favored by a propitious international and regional environment, and by the elites’ abandonment of policy radicalism and embracement of democratic values (Mainwaring and Pérez-­Liñán 2013; Gervasoni 2014). The abysmal performance of the 1976–83 military government was critical in making Argentine voters and leaders more moderate and democratic. Even the revolutionary armed left espoused democracy after suffering the traumas of military defeat and state terrorism in the second half of the 1970s (Ollier 2009). Argentina is a federal democracy with three tiers of government: national, provincial, and municipal. Provinces, which are constitutionally mandated to sustain democratically elected governors and legislative bodies, are power­ ful and quite autonomous (Hooghe et  al. 2016). Figure I.3 presents a map of Argentina and its twenty-­three provinces (plus the Autonomous City of Buenos Aires), identifying with light gray those that are reasonably democratic according to all the measures used later in the book, with dark gray those that seem imperfectly democratic under most measures, and with an intermediate

 Only one case in their sample of Latin American countries, Peru, had more transitions (seventeen).

12

Introduction

16

JUJ FSA

STA CAT

TUC

SE

CHA

MIS CTS

LR SJ

SF CBA

MZA

ER CF

SL

LP

PBA

NEU RN

CHU

SC

TF

Regime type Democratic

Intermediate

Less democratic

Figure I.3.  Argentina’s provinces and their level of democracy (1983–2015)

shade of gray those that are in between (or for which different measures disagree).13 This necessarily simplified three-­color scheme should be taken as a first approach to the matter. Chapters 2 and 3 provide statistics and two additional maps that allow for a more nuanced differentiation of levels of democracy among provinces and within provinces over time.  The Islas Malvinas/Falkland Islands (shown in the map), Islas Georgias del Sur/South Georgia Islands (not shown), and Islas Sandwich del Sur/South Sandwich Islands (not shown), as well as their corresponding maritime spaces, are claimed by Argentina but are under de facto colonial administration of the United Kingdom. Because these islands are legally part of the province of Tierra del Fuego, Antártida e Islas del Atlántico Sur, the Islas Malvinas/Falkland Islands are painted with the same color as that province.

13

Introduction

17

According to the Regional Authority Index (Hooghe et  al. 2016), the Argentine provinces are among the most authoritative subnational units in the world,14 not far below the German länder and the Australian and US states, at par with the Canadian provinces, and somewhat ahead of the Mexican and Brazilian states. Provinces write their own constitutions and choose their own electoral rules. In some of them, for example, the governor cannot be reelected immediately, while in others there are no term limits (Almaraz 2010; Cardarello 2012). Argentine provinces have primary responsibility over key policy areas such as health services, primary and secondary education, and public safety (including their own police forces). This means that they account for a significant share of public sector expenditures (around 36–40 percent of all Argentine public spending in the period covered by this book), and for about two-thirds of all of Argentina’s public employees. Provincial governments collect some taxes, but most of their fiscal revenues originate in transfers from the central government, which levy the country’s main taxes: the VAT, the income tax, and taxes on financial transactions, imports and exports (for a more detailed explanation of the functioning of fiscal federalism in Argentina, see the end of this section and Section 5.2). Provinces are represented in the national Senate at a rate of three popularly elected legislators each (two by the winning party, one by the runner up), so that all national legislation is formally subjected to approval by representatives of the provinces. Argentina’s bicameralism is symmetrical: the Senate has roughly the same powers to legislate and control the executive branch as the Chamber of Deputies (Llanos and Nolte 2003). Provinces are key players in national politics not just because of their representation in the Senate. They also constitute the districts for the election of deputies, so all national legislators are in a sense representatives of the provinces. Governors command significant administrative, fiscal, and patronage resources. Presidents often negotiate directly with them, as their cooperation is important to pass legislation through congress (because legislators from their parties and provinces are often subordinated to them15) and to implement national policies in the provinces. Not surprisingly, most presidents and viable presidential candidates during the current democratic period had previously been governors. A key feature of Argentina’s federal and legislative politics is malapportionment. Demographically small provinces are strongly overrepresented in the Senate (a common feature of federations) but also significantly overrepresented in the lower chamber (Samuels and Snyder 2001). This is due both to constitutional (for the Senate) and legal (for the Chamber of Deputies) design, and to the very uneven territorial distribution of Argentines: the province of Buenos Aires alone contains 39 percent of the population, a figure that rises to 62 percent when the next three largest districts are added (Ciudad Autónoma de  Data for the most recent year available, 2010.  See Gervasoni and Nazareno (2017).

14 15

18

Introduction

Buenos Aires, Córdoba and Santa Fe). At the other end, the twelve demographically smaller provinces (most of them in the North of the country and in the southern region of Patagonia) account for only 13.6 percent of the population (but control half of the senators).16 This legislative and demographic configuration gives “low-­maintenance” provinces (Gibson and Calvo 2000) – those with small populations and a disproportionate number of national senators and deputies – significant political clout. The rentier theory of subnational democracy I summarized in the previous section assigns much explanatory power to the way polities are financed. Subnational units can obtain their fiscal revenues either from their own taxes, or from transfers from the central government. Up to the early 1930s, the national government and the provinces had their own, separate, sources of tax revenues, which meant that there were no “fiscal federalism rents”: provinces did not receive on a regular basis transfers from the central government. The Great Depression hit hard the main source of national revenues – tariffs on foreign trade – and pushed authorities to create sales, income, and excise taxes. These reforms implied that the national government centralized taxes that were previously levied by the provinces (excise taxes) or exploited tax bases that were legally under (at least partial) provincial jurisdiction (sales and income taxes). To compensate provinces for these losses, each new national tax law contained provisions to share part of the revenues with the provincial governments. Thus, intergovernmental revenue-­sharing was born in Argentina in 1935. Changes in the institutions of fiscal federalism since then have been frequent and haphazard, a fact partly explained by the regime instability described earlier (on the historical evolution of fiscal federalism in Argentina, see Eaton 2001; Cetrángolo and Jiménez 2004; Porto 2004; Díaz-­Cayeros 2006, Chapter  7). Some general trends, however, can be recognized. The system expanded over time to cover new taxes and new territories (there were only fourteen provinces in 1935; there are twenty-­three now, plus the autonomous city of Buenos Aires17), and it tended to increase the share of revenues corresponding to the provinces, at the expense of the federal treasury (Cetrángolo and Jiménez 2004). The initial devolutive criteria that assigned revenues to provinces in proportion to their tax contribution, population or public spending, was in successive reforms replaced by more redistributive criteria. Although the original goal was to favor the least developed regions (Díaz-­Cayeros 2006), over time the new laws tended to privilege demographically small provinces (underdeveloped or not) at the expense of the larger districts, and in particular of  Figures based on official figures from the 2010 Argentine Census (www.indec.gov.ar).  Buenos Aires, Catamarca, Córdoba, Corrientes, Entre Ríos, Jujuy, La Rioja, Mendoza, Salta, San Juan, San Luis, Santa Fe, Santiago del Estero, and Tucumán. Nine new provinces were created in 1951 (Chaco and La Pampa), 1953 (Misiones), 1955 (Chubut, Formosa, Neuquén, and Río Negro), 1956 (Santa Cruz) and 1990 (Tierra del Fuego, Antártida e Islas del Atlántico Sur). Finally, the city of Buenos Aires acquired a province-­like status of autonomy in 1996.

16 17

Introduction

19

the largest, the province of Buenos Aires. By the 1980s, when democratic politics was permanently reestablished both at the national and provincial level, the system had already acquired the characteristics that create the conditions for subnational rentierism: (1) provinces are in charge of a significant share of all public sector spending; (2) provincial expenditures are mostly financed by federal transfers (or, more technically, there is a significant vertical fiscal imbalance); and (3) demographically smaller provinces are strongly favored by the revenue-­sharing system. These characteristics imply that some provincial governments have been heavily subsidized during the 1983–2015 democratic period covered by this book, allowing them to spend amounts of money that would be unthinkable if they had to be financed by local taxpayers (Section 5.2 provides a more elaborate account of this rentier strain of fiscal federalism).

I.5  Plan of the Book Part I of the book (“Description: The Anatomy and Evolution of Subnational Regimes”) provides a detailed account of democracy and authoritarianism in the provinces of Argentina. Chapter 1 operationalizes the concept of subnational democracy. After defining it and disaggregating it into many dimensions and components, I propose: (1) an objective measurement strategy based on electoral and institutional indicators (which results in the Subnational Democracy Index); and (2) an alternative subjective strategy based on the SEPP (which produces several indices corresponding to different aspects of subnational regimes). The next two chapters are dedicated to fleshing out each measurement strategy. After introducing the methodological details of the objective Subnational Democracy Index, Chapter 2 uses it to describe differences in levels of democracy across provinces and over time. To complement these quantitative results, the chapter ends with a qualitative account of the typical institutions and practices that characterize the less democratic (or “hybrid”) provinces and differentiate them from their more democratic counterparts. Chapter 3 focuses on the second measurement strategy, the SEPP. Because this survey of experts is in many ways a novel methodology to evaluate subnational democracy, its design is explained and justified in detail. Most of the rest of the chapter uses the rich SEPP data to provide a comprehensive description of many dimensions of provincial regimes in Argentina. In the final section, the results of the objective and subjective measures are compared for the period for which they overlap (2003–7) to assess to what extent they produce similar estimates of the level of democracy in each province. Part II (“Explanation: The Causes of Subnational Regimes”) is about the factors that may plausibly account for wide interprovincial differences in levels of democracy. In Chapter  4, I briefly review existing theories of subnational regimes, and discuss the state-­society balance, fiscal, and rentier-­state approaches that constitute the foundation of the rentier theory of subnational democracy I put forward. The chapter also proposes and justifies a rigorous

20

Introduction

definition of the concept of “rent” that, by occupying a relatively high rung on the ladder of abstraction, encompasses resource rents, fiscal federalism rents, and other types of rentier income. Following a causal understanding of concepts, at the end of the chapter I discuss the causal pathways from rents to regimes suggested by the proposed definition. Chapter  5 develops the main causal argument  – the rentier theory of subnational democracy  – at length, explains how Argentina’s intergovernmental revenue-­sharing rules generate large fiscal federalism rents for several provinces, and expounds how in these subnational rentier states democratic institutions are undermined by authoritarian practices, giving rise to hybrid provincial regimes. The chapter also spells out and justifies the specific causal mechanisms expected to connect fiscal federalism rents with lower levels of provincial democracy. Chapter 6 subjects the rentier theory of subnational democracy to several statistical tests, with an emphasis on showing that the key finding – that fiscal federalism rents are detrimental to provincial democracy – is robust to a great variety of measurement, specification, and estimation choices. Its last section presents qualitative evidence from several provinces on the causal mechanisms theorized in the previous chapter, emphasizing as the key link the high levels of economic dependence on the provincial budget that affect most individuals, groups, companies and media outlets in rentier provinces. Part III of the book (“Comparison: Subnational Regimes around the World”) places Argentina in comparative perspective. In Chapter 7, I introduce the objective CDSI and test it on the subnational units of seven federations (Argentina, Australia, Canada, Germany, India, Mexico, and the United States) and one unitary country (Uruguay). The chapter contains both a methodological contribution – an index of subnational democracy that can be easily applied to the diverse institutional settings of subnational units in different countries – and a substantive contribution – the first study (to my knowledge) comparing levels of subnational democracy in all the regions of several countries. In the final chapter I conclude, summarizing the book’s main descriptive and causal claims, discussing their theoretical, normative, and policy implications, and suggesting important areas for future research on subnational regimes.

Part I DESCRIPTION: THE ANATOMY AND EVOLUTION OF SUBNATIONAL REGIMES

1 Defining and Measuring Subnational Regimes*

As scholars of democracy turn their sights to subnational regimes, a first and basic challenge is that of conceptualization and measurement. How do we determine whether a province is democratic, authoritarian, or “hybrid”? The pioneering studies on “subnational authoritarianism” (Cornelius 1999; Snyder 1999; Gibson 2005) helped put the topic on the research agenda and provided valuable descriptions of some of the least democratic subnational units in national democracies around the world, but they generally did not provide an explicit conceptual definition and set of indicators needed to classify these and other units as authoritarian. There is, however, a long and solid literature on the operationalization of regime types at the national level. Indices of democracy covering most countries in the world have been available for decades (e.g., Bollen 1980; Coppedge and Reinecke 1991; Alvarez et  al. 1996; Vanhanen 2000; Marshall, Gurr, and Jaggers 2014; Freedom House 2016), as well as several region-­specific indices (for Latin America see Mainwaring, Brinks, and Pérez-­Liñán 2001; Bowman, Lehoucq, and Mahoney 2005; Mainwaring and Pérez-­Liñán 2013). The measurement of national regimes has given rise to a sophisticated methodological debate (Bollen 1993; Bollen and Paxton 2000; Munck and Verkuilen 2002; Goertz 2006; Treier and Jackman 2008; Munck 2009; Coppedge 2012) that has led to a better understanding of the advantages and disadvantages of the extant indices, and has provided guidance on how to design measures that maximize validity and reliability. Since 2000, several works have proposed novel cross-­national indices that – using new conceptualizations, new data and/or new statistical methods – improve on the previous ones (Paxton 2000; Coppedge, Alvarez and Maldonado 2008; Treier and Jackman 2008; Pemstein, Meserve and Melton 2010; Coppedge and Gerring et  al. 2011; *

  Portions of this chapter were published previously in Gervasoni (2010b, 2016a).

23

24

Description: The Anatomy and Evolution of Subnational Regimes

Boix, Miller, and Rosato 2013). The young subfield of subnational regimes is, by comparison, light-­ years behind in conceptual clarity, measurement rigorousness, and data richness. The earliest studies I am aware of which define and measure democracy systematically in all (or most of) the regions of a country are those conducted by Hill (1994) for the United States, Hernández Valdez (2000) for Mexico, and McMann and Petrov (2000) for a large subset of the regions of Russia. More recently, scholars have used objective institutional or electoral indicators to construct indices that measure subnational democracy or subnational electoral contestation (an important dimension of democracy) in countries such as Argentina, Brazil, India, Mexico, and the United States (Beer and Mitchell 2006; Borges 2007; Goldberg, Wibbels, and Mvukiyehe 2008; Gervasoni 2010a; Giraudy 2010, 2015; for an exception using subjective indicators for the cases of Russia and Kyrgyzstan, see McMann 2006; for an index that combines both types of indicators, see Saikkonen 2016a1). These operationalization efforts go in the right direction, but still pale in comparison to their national counterparts. Given this state of affairs, before tackling the problem of explaining variance in subnational regimes (Chapters 4–6), this book dedicates considerable space to operationalizing its dependent variable  – the degree of subnational democracy2 – and to providing a comprehensive description of the Argentine provinces in terms of this variable (Chapters 2 and 3). Chapter 7 compares the level of subnational democracy in eight countries. In terms of conceptualization I make two consequential decisions: (1) focusing on the “level of democracy” rather than on the “quality of democracy,” and (2) using a thick and multidimensional definition of democracy.3 With respect to measurement approaches, I draw on the two existing national traditions – objective and subjective – to design two alternative instruments to assess provincial regimes in Argentina, plus one more general objective index that I apply to several comparative cases (Chapter 7). Using five objective indicators, I developed the Subnational Democracy Index and applied it to all the Argentine provinces from 1983 to 2015 (Chapter 2). I also developed several subjective indices of different aspects of subnational regimes on the basis of the data produced by the Survey of Experts on Provincial Politics (SEPP) which I carried out during 2008 (Chapter 3). This operational strategy is much more costly in terms of resources and time, but is likely more valid and surely thicker  Saikkonen uses an index that combines objective electoral data with an unpublished expert rating of Russia’s regions compiled by Petrov and Titkov (2013). 2  I do not use the expression “subnational regime type” because it implies that there are only two or a few types, such as democracy and authoritarianism. As the rest of this chapter makes clear, regimes in general (and subnational regimes in particular) are more appropriately positioned on a continuum that goes from more to less democratic than in discrete categories. 3  Few works have focused on the conceptualization of subnational regimes; for an exception see Giraudy 2013. 1

Defining and Measuring Subnational Regimes

25

than the objective approach. Many different aspects of democracy were measured, which resulted in a large number of indicators and indices. The temporal scope of the survey, however, is limited to the period to which the questions refer (2003−7). As objective and subjective indicators possess complementary strengths and weaknesses, using both to look into the provincial regimes of Argentina allows for a deeper, fuller, and better understanding of them.

1.1  The Background Concept: Democracy Following Adcock and Collier (2001) in this and the next two sections I review the first three levels in the definition and measurement of concepts: the background concept, the systematized concept (including its disaggregation into dimensions and subdimensions), and the indicators.4 Few concepts have been more politically contested than democracy. In the second half of the twentieth-­century the word “democracy” became so prestigious and legitimizing that all types of political philosophies and regimes tried to appropriate it. One or another “model of democracy” has been advocated both by a long tradition of liberal thought – from John Locke, Baron de Montesquieu, and James Madison to Friedrich Hayek and Robert Nozick − and by the later socialist school, from Karl Marx to Nicos Poulantzas and Crawford B. Macpherson (Held 1987). Likewise, both capitalist countries with multiparty elections and communist nations ruled by single parties have claimed to embody the principles of democracy. Influential and well-­founded conceptions of democracy give priority to significantly different values, for example popular rule, prevention of tyranny, human development, and political community (Katz 1997). Moreover, alternative democratic values are often at odds with each other, as illustrated by the clear tension between the principles of majority rule and minority rights. To derive a clear, systematized concept of democracy from this noisy background I follow two guidelines: (1) taking into account its etymological meaning, and (2) considering the sense in which the word is generally used in influential social sectors outside academia (for example in the realm of practical politics, the media, etc.).5 Democracy was originally conceived by the ancient Greeks as a type of krátos, or rule, and still today most citizens, politicians, and academics think of it as a characteristic of governments or regimes (as opposed to the society or the economy). The types of krátos that can be characterized as democratic are those based on the demos, the people.

 Adcock and Collier’s fourth and last level, the scoring of the cases, is explained in Sections 2.1 (where the objective Subnational Democracy Index is explained in detail) and 3.1 (where the methodology of the Survey of Experts on Provincial Politics is presented). 5  These criteria privilege “resonance” (Gerring 2001, 52–4), which in turn facilitates communications with people who are not scholars (like policy makers and journalists). 4

26

Description: The Anatomy and Evolution of Subnational Regimes

In its travel from classical Athens to the contemporary world, democracy picked up two important elements. The first, originating in its marriage with liberal ideas, was freedom. The second, emerging from the impossibility of applying direct democracy to large states, was representation. Therefore, I understand democracy as a political regime in which rulers are periodically chosen in competitive elections by the people and, once in office, exercise power in a limited way, respecting political rights and civil freedoms. This conceptualization avoids the extremes of maximalist and minimalist definitions (Munck and Verkuilen 2002, 9−12): it does not include attributes beyond a strictly political concept of democracy, but it does not leave out the liberal element, which I consider critical while some prominent political definitions do not.6 From this point of view, a regime is undemocratic (or authoritarian) when it does not conduct elections (or conducts uncompetitive, unfair, or exclusionary elections) or when, even if headed by popularly elected leaders, the power of the state is used to kill, incarcerate, exile, or otherwise punish citizens for political reasons. This political definition means that democracy is a characteristic of national, subnational and supranational regimes, not of the economy, the society, or the mass culture. Democracy can (and does) coexist with high income inequality, hierarchical social organizations, and politically apathetic citizens. None of these is desirable, but none of them makes a country less democratic. Cuba is economically more egalitarian than Brazil, but infinitely less democratic. Good things do not always go empirically together. Putting them conceptually together does not help improve the world, and it most certainly hinders our attempts to understand it. 1.1.1  Level or Quality of Democracy? An alternative “background concept,” that of “quality of democracy” (Diamond and Morlino 2004), has vigorously emerged in the last 20 years, in part as a way of dealing with new democracies that conform to standard definitions, but remain somehow unsatisfactory from a normative point of view. I find three problems that make the use of this concept unadvisable for my goals. First, there is no agreement on whether the quality and the degree of democracy are the same thing. Some authors posit that “quality” is different from “degree” (e.g., Altman and Pérez-­Liñán 2002, 87), while others use those words interchangeably (e.g., Lijphart 1999, 276). Second, even accepting that quality is different from degree, it is often unclear whether the term “quality” refers only to democracy, or both to democracy and to governance (Plattner 2004). In the latter case, the concept becomes too wide and incoherent as there 6

 Joseph A. Schumpeter (1975 [1942]) and Alvarez et  al. (1996), for example, also propose strictly political definitions, but they emphasize electoral contestation, downplaying the liberal dimension.

Defining and Measuring Subnational Regimes

27

are democracies with good and bad governance, and autocracies with good and bad governance. Third, standard definitions of the “quality of democracy” often extend their reach to areas that are strictly speaking not political, in the sense of not being directly related to the government or the regime. This is well illustrated by items included in audits of democratic quality, such as “a climate of opinion that rejects all types of bigotry and discrimination,” the proportion of “individuals who are unemployed” (O’Donnell 2004, 43 and 63), or the “democraticness” of the political culture (Proyecto Estado de la Nación 2001, 31). In summary, the young concept of quality of democracy seems to be still too fuzzy, broad and/or demanding to be useful for the goals of this book. I focus on a different variable, the level of (subnational) democracy. The next section defines this variable.

1.2  The Systematized Concept: Liberal Representative Democracy In this section, I define the “systematized concept” (Adcock and Collier 2001) that I will use in the rest of the book. I call it here “liberal representative democracy,” which is a type of political regime in both national and subnational polities. The definition I choose is relatively thick,7 complex, and multidimensional. Even in a strictly political sense, liberal representative democracy has several dimensions and subdimensions. For example, the specifically democratic component  – popular sovereignty  – and the liberal component  – individual rights and limited power – are conceptually different, historically independent, and sometimes empirically uncorrelated (as is the case in “illiberal democracies”, see Zakaria 1997). This “thick” conception contrasts with an important “thin,” Schumpeterian8 definitional tradition that deliberately wants “to define ‘democracy’ narrowly” as “a regime in which those who govern are selected through contested elections” (Przeworski et al. 2000, 15). Finally, I understand democracy as a continuous variable (as opposed to a dichotomous or ordinal one). Both for theoretical and empirical reasons, I follow those who think of democracy in terms of levels or degrees (Coppedge and Reinicke 1991), including a well-­established tradition of measurement of democracy.9 However, authors who advocate a dichotomous definition (e.g., Alvarez et  al. 1996), make an important point: a regime  – however

 I follow Coppedge’s (1999) recommendation to combine “thick” conceptualizations, typical of the qualitative tradition, with a rigorous operationalization amenable to subsequent quantitative analysis. 8  See note 6. 9  Such as the indices proposed by Freedom House 2016, the Polity IV project (Marshall, Gurr, and Jaggers 2014), Vanhanen 2000, and the Varieties of Democracy project (Coppedge et  al. 2018a, b). 7

28

Description: The Anatomy and Evolution of Subnational Regimes

liberal  – cannot be called democratic if rulers are not chosen in competitive elections. Therefore, democracy here is conceived as a largely continuous concept once the dimension of contestation reaches a minimum threshold: the existence of meaningful multiparty popular elections. This threshold sharply divides polities with elective rulers from monarchies, military dictatorships, single-­party totalitarianisms, and other types of regimes in which decision-­ makers are not chosen through competitive elections. Democracy is a type of political regime, which in turn is defined by a set of (formal or informal) rules that determine the type of actors who can occupy the main positions of government, the accepted methods to obtain those positions, and the way in which public-­policy decisions are made (Schmitter and Karl 1991, 77; Munck 1996). More succinctly, a regime is a set of rules that regulate: (1) how government positions are filled, and (2) what government officials can and cannot do. In the case of democracy, the first set is anchored in the principle of what Katz (1997) calls popular sovereignty – which finds institutional expression in popular, competitive, free, and fair elections – while the second derives from the principle of limited government – implemented, for example, through the institutions of separation of powers, checks and balances, and constitutionally protected individual rights  – akin to Katz’s “tyranny prevention” principle, to Held’s (1987) “protective democracy,” and to Riker’s (1982) liberal justification of democracy. Efforts to measure national democracy have yielded similar dimensions: Bollen and Paxton (2000, 59−60), for example, argue that liberal democracy is a political system characterized by both “democratic rule” and “political liberties.” In summary, these dimensions correspond to the strictly “democratic” and the “liberal” dimensions of modern representative regimes. Liberal democracy, then, is defined by the combination of competitive elections and limited power. If rulers are competitively elected but exercise power with few limitations, the regime is an illiberal democracy (Zakaria 1997; Diamond 1999, 42−49) or a democradura (O’Donnell and Schmitter 1986, 9). Alternatively, unelected rulers may exercise power in a limited fashion, which results in liberal autocracies (Diamond 1999, 4) or dictablandas (O’Donnell and Schmitter 1986, 9). Finally, when rulers are not elected and power is exercised without limits, the regime is authoritarian.10 The foregoing discussion leads to a first general definition of liberal democracy: Liberal democracy is a type of political regime in which the top government positions are filled directly or indirectly through contested elections, and in which government power is divided among different branches that check each other, and limited by

 Using the word “authoritarian” in a broad sense, including both Linz’s (1975) authoritarian and totalitarian regimes.

10

Defining and Measuring Subnational Regimes

29

constitutionally or legally mandated (and de facto respected) political and civil liberal rights and freedoms.

Once the concept of liberal representative democracy is fleshed out into dimensions and subdimensions (see Section 1.2.2), this definition can be expanded into a thicker and more complex one by adding the details that go into the definition of each of these dimensions. 1.2.1  The Other End: Authoritarianism or Hybrid Regimes? The last decade of the twentieth century saw the rise of a great number of regimes that cannot be easily classified as either authoritarian or democratic but display some characteristics of both. (Ottaway 2003, 3)

Although the theoretical range of the variable regime type goes from democratic to authoritarian, the real range in given empirical domains may be narrower. Despite the relatively common use of concepts such as “authoritarian enclaves” and “subnational authoritarianisms” to describe some regional regimes in federal democracies (Fox 1994; Cornelius 1999; Diamond 1999; Gibson 2005; Gel’man and Ross 2010), even the least democratic Argentine provinces do not meet the conventional definition of authoritarianism, and the same seems to be true in other third-­wave democratic federations, such as Brazil, Mexico, and Russia during the 1995–2004 period, when regional governors were elected. Subnational units in these contexts are generally far from being the kind of repressive and closed regimes that the Polity IV database codes as “autocracies” and Freedom House labels as “not free.” These regimes have elections (often reasonably free), real opposition parties, minority representation in the legislature, nontrivial levels of freedom of speech, and so forth. One does not find in the Argentine provinces bans on political parties, incarcerated dissidents, or significant media censorship. Because they are embedded in a national democracy, subnational leaders are constrained in the extent to which they can restrict political rights. Given that at the national level democracy is widely accepted as “the only game in town,” and that the Constitution empowers national authorities to guarantee democracy in the provinces,11 there are strong incentives for self-­interested provincial rulers to avoid blatantly authoritarian practices such as jailing opposition leaders or massively 11

 The most powerful instrument in this respect is the removal of provincial authorities via federal intervention, an attribution given by the national Constitution to congress and the president. Sustaining democratic institutions (“guaranteeing the republican form of government,” in the nineteenth-­century language of the Constitution) is one of the few legal justifications for interventions. Since 1983 four provinces were intervened six times: Catamarca (1991), Tucumán (1991), Corrientes (1991 and 2000), and Santiago del Estero (1993 and 2004). All of them except Tucumán had doubtful democratic credentials.

30

Description: The Anatomy and Evolution of Subnational Regimes

rigging elections. Such visible actions easily attract attention from the national media, hurting the chances of the perpetrators in national politics and increasing the likelihood that national authorities will take corrective measures. The less-­democratic provincial regimes, then, combine democratic institutions that are not just a façade with practices that are clearly, if subtly, authoritarian. They are well conceptualized by the literature on (national) hybrid regimes (Karl 1995). The definitional traits of “semi-­democracies” (Diamond, Linz and Lipset 1995; Mainwaring, Brinks and Pérez-­Liñán 2001), “illiberal democracies” (Zakaria 1997), “competitive authoritarianisms” (Levitsky and Way 2002, 2010), “semi-­authoritarianisms” (Ottaway 2003), and “electoral authoritarianisms” (Schedler 2006) describe the less-­democratic Argentine provinces more accurately than the traditional concept of “authoritarianism” (Linz 1975). Moreover, the causal logic at work seems to be similar: just as national hybrid regimes exist largely because of the need to avoid overt authoritarianism in the face of strong international pressures for democratization, subnational leaders with authoritarian projects come under intense national pressure to sustain at least minimal levels of democracy. For these reasons, the concept of hybrid regimes (or related concepts such as “electoral authoritarianism”) has been used in the past decade to characterize the least democratic regions of several third-­wave federations (McMann 2006; Gervasoni 2010a; Borges 2016; Saikkonen 2016a). Whether a subnational unit qualifies as authoritarian or hybrid is, of course, an empirical question. The measurement strategies outlined in Section 1.3 are able to identify all types of regimes, from roundly democratic to clearly authoritarian. The evidence presented in Chapters  2 and 3, however, supports the idea that the actual empirical range of contemporary provincial regimes in Argentina runs from democratic to hybrid. Therefore, I refrain from using the concept of subnational “authoritarianism” to characterize the least democratic provinces, using the expression “hybrid regimes” instead. They surely contain elements of authoritarianism, but these elements do not necessarily make them authoritarian given that they are combined with elements of democracy, and in particular with real multiparty elections. 1.2.2  The Dimensions and Subdimensions of the Concept The proposed definitional perspective decomposes the concept of democracy in two dimensions12, a strictly democratic one (Katz’s “popular sovereignty”), and a liberal one (“limited government”). These dimensions are conceptually different, even if theoretically and empirically related (Diamond 1999, 4–5).

 What I call dimensions and subdimensions corresponds to what Munck and Verkuilen (2002) call “attributes” and “components,” respectively.

12

Defining and Measuring Subnational Regimes

31

1.2.2.1  The Democratic Dimension: Contested, Inclusive, and Effective Elections This dimension expresses the “democratic” side of “liberal democracy.” Drawing on Dahl (1971), Hadenius (1992, 49–51), and Diamond, Linz, and Lipset (1995, 6–7) I identify three subdimensions: contestation, inclusiveness (Dahl’s original dimensions), and effective elections. I define them, conventionally, as follows: Contestation.  The extent to which individuals and groups can effectively oppose the incumbent authorities and participate in regular elections that are competitive and, therefore, reasonably likely to lead to the defeat of the incumbent. Once the minimum threshold of contestation is achieved (i.e., once rulers are chosen in reasonably free multiparty elections), there may be higher or lower levels of contestation depending on many and diverse factors, such as the level of barriers to entry of political parties, the rules of campaign financing, the degree of media plurality, the prevalence of patronage, and so forth. Inclusiveness.  The proportion of the adult citizenry who enjoy, legally and factually, the rights associated with political competition, especially the rights to vote and run for office. Given current democratic standards, a polity is considered fully inclusive only if all, or practically all, adult citizens enjoy these rights. Effective Elections.  A key condition for democracy is that elected officials are not subordinated to unelected ones (Schmitter and Karl 1991, 81). Elections are not “effective” if “elected organs are limited in their decision-­ making by instances which, for their part, have no democratic support” (Hadenius 1992, 49). There are several polities in the world in which elected officials have to yield, at least in some policy areas, to monarchs, generals, theocratic elites, or unelected legislators. 1.2.2.2  The Liberal Dimension: Institutional Constraints and Individual Rights Liberalism is grounded in the tradition of Western thought that emphasizes individual rights (Locke), institutional constraints on state power (Montesquieu), limits to state intervention in people’s lives (Constant), and protection against the tyranny of the majority (Hamilton, Madison, and Jay; J. S. Mill). The overall principle is to protect freedoms, both those strictly necessary for effective contestation, usually known as political rights (freedom of organization, assembly, speech, etc.), and those not strictly political but still important for individual autonomy, usually known as civil liberties (freedom to choose place of residence, freedom of movement, religious freedom, and so forth).13

 Even in attempts at measuring democracy that do not include indicators of liberal rights, there is a recognition that they are part of the definition (Marshall, Gurr, and Jaggers 2014, 14) or that they have been historically associated with the concept of democracy (Alvarez et al. 1996, 4).

13

32

Description: The Anatomy and Evolution of Subnational Regimes

How are these freedoms protected? The answer to this question defines the two subdimensions of the liberal dimension of democracy: Institutional Constraints.  Classic liberals proposed the separation of powers, so that different aspects of government would be in different hands. Montesquieu’s idea of “checks and balances” finds institutional expression in all contemporary democracies which, to different extents, separate executive, legislative, and judicial functions in different bodies. Moreover, additional “agencies of horizontal accountability” (O’Donnell 1999a) operate in democracies to control and limit the excesses of the executive (e.g., anticorruption offices). This dimension is typically given less consideration than competitive elections and political rights in existing conceptualizations and operationalizations of democracy.14 Individual Liberal Rights.  A second liberal artifact to limit state power and maximize individual freedom is the legal protection of political rights and civil liberties. Explicit and clear laws about the freedoms that are granted to individuals make it more costly for the state to violate them, and easier for the courts to protect them. These rights can be seen as a third dimension implicit in Dahl’s two classic ones (Diamond, Linz and Lipset 1995, 6–7). Of course, the fact that these freedoms are written in constitutions or statutes and formally protected by courts does not mean that they are factually respected. The extent to which they are, then, is a central component of the liberal dimension of democracy. The five proposed subdimensions resemble those defined by some mainstream conceptualizations in political science. For example, Mainwaring, Brinks, and Pérez-­Liñán (2001) identify four characteristics of democracy: (1) free and fair competitive elections, (2) inclusive adult citizenship, (3) civil liberties and political rights, and (4) actual governing by elected authorities. These correspond almost perfectly to my contestation, inclusiveness, liberal rights, and effective elections subdimensions, respectively. The main difference is that I include a fifth subdimension, institutional constraints. Figure 1.1 presents the structure of the concept of democracy elaborated so far, going from the highest level of abstraction (the genus) to the lowest (the subdimensions). 1.2.2.3  Components and Subcomponents of Democracy The “thick” conceptualization of political regimes adopted in this book implies that there are many specific aspects of democracy into which each subdimension  The Polity IV indicator “executive constraints” or “institutionalized constraints on the decision-­ making powers of chief executives” (Marshall, Gurr, and Jaggers 2014, 24) is a very important exception. Interestingly, this indicator is highly correlated with the Bollen, Freedom House and Vanhanen indices of democracy, and it “virtually determines the democracy and autocracy scale values” of the Polity scores (Gleditsch and Ward 1997, 379−80).

14

Defining and Measuring Subnational Regimes Broader category (genus)

Political regime

Liberal representative democracy

Systematized concept (differentia)

Popular sovereignty

Dimensions (or attributes) Subdimensions (or components)

33

Inclusiveness

Effective elections

Political liberties

Institutional constraints

Liberal rights

Contestation

Figure 1.1.  Genus, differentia, dimensions, and subdimensions of democracy

of democracy is disaggregated. Table 1.1 is a systematization of the dimensions (column 1) and subdimensions (column 2) defined so far, plus the components (column 3) and subcomponents (column 4) of each subdimension. Columns 1 through 4 should be interpreted as general categories applicable to the elective subnational regimes of any polity, while the indicators used to measure each of them have to be adapted to the context – in this case to the Argentine provinces – and to each particular province (e.g., survey items will be different for “presidential” and “parliamentary” regions). As the Comparative Subnational Democracy Index I introduce in Chapter 7 and the replication of the SEPP in Mexico15 illustrate, some of the indicators proposed below for Argentina may be directly applicable to other countries, while others need to be adapted to their different institutional contexts. Each of the twenty-­two subcomponents listed in the last column of Table 1.1 is measured through one or more of the items (questions) contained in the SEPP (see Chapter 3). The objective measurement strategy cannot produce a similarly disaggregated set of indicators, so it will only operate at the level of the two main dimensions. The next section provides a detailed explanation of both types of indicators.

1.3  Indicators: Objective and Subjective Measures of Democracy In this section, I go one level below the “systematized concept” to address the indicators that will be used to measure the concept of (subnational) liberal representative democracy.

 See Loza and Méndez 2016.

15

34

Description: The Anatomy and Evolution of Subnational Regimes

Table 1.1.  Disaggregation of the concept of subnational democracy Dimensions

Subdimensions

Popular sovereignty Inclusiveness or democratic rule Contestation

Components

Subcomponents

Extension of effective right to vote Extension of effective right to run Fairness of elections

Denial of right to vote

Freedom of expression

Effective elections Political liberties or limited government

Institutional constraints

Liberal rights

Denial of right to run Fairness of campaign Fairness of electoral act and vote counting Opposition leaders Critical journalists Politically relevant media Public employees General population Political parties

Freedom to form/join organizations Unelected local powers Unelected local powers Elected national Elected national powers powers Legislature Provincial legislature Judiciary Provincial justice Agencies of Horizontal Independence of Accountability agencies of HA Incumbent Party Constraints of party on governor Freedom of Right to alternative information and diverse sources of information Access to information about government Personal freedoms Physical security Privacy Alternative or minority lifestyles Academic freedom

Defining and Measuring Subnational Regimes

35

For several reasons, subnational regimes pose harder measurement challenges than national regimes. First, there are decades of experience in developing national indices of democracy. Over time, these indices have been subjected to scrutiny and critiques, and to the pressure of competition from alternative ones. As a result, newer and sounder ones have superseded older and weaker measures, and the methodological standards of the surviving ones have improved significantly.16 Subnational measures of democracy are fewer, more recent, and have been applied only to a few countries. The prevalence of hybrid regimes in subnational settings (see Section 1.2.1) pose two additional measurement problems. First, the range of regime variation is narrower at the regional than at the national level  – because many fully autocratic national regimes do exist  – which leads to weaker reliability.17 Second, the mix of elements of democracy and authoritarianism that defines hybrid regimes inevitably increases measurement error: when a regime is homogenous, all indicators will yield similar scores; conversely, when the regime is heterogeneous, different indicators will yield different scores. Saudi Arabia will be coded as authoritarian regardless of the measure used, while Indonesia or Turkey will appear more democratic according to some indicators and less so according to others. The fact that the least democratic subnational regimes in Argentina and other countries are hybrid (as opposed to authoritarian) conspires against reliability. Second, the quantity and quality of secondary sources to assess regional regimes is generally lower than that available to evaluate national regimes. Even the most important provinces of Argentina have been subject to few rigorous academic descriptions, and the information in the media about many of them is scarce, low quality, or biased.18 This source of measurement error is especially relevant in smaller and less developed provinces, where secondary sources are especially wanting. Figure 1.2 presents these measurement challenges graphically. The horizontal line represents the total range of regime variation, from fully autocratic to fully democratic (with four countries illustrating each pole). The gap in the line reflects the discrete threshold between regimes that do not hold multiparty elections and electoral regimes. A democratic province (Mendoza) and a less democratic one (Formosa) are placed on this continuum. The distance between them is considerably smaller than that between Cuba and Canada. The circles and triangles represent scores from repeated measurements (or from different

 The Varieties of Democracy project is the clearest example of such improvement. See Coppedge and Gerring, et al. 2011, Coppedge et al. 2018a, b, and https://v-dem.net/en/. 17  Reliability depends not only on the precision of the measurement instrument but also on the magnitude of the differences among the objects being measured (Traub 1994). A given level of measurement error may be small for objects that are far apart, but too great to distinguish objects that are closer to each other. 18  As the indicators on press freedom in Chapter 3 will make abundantly clear. 16

Description: The Anatomy and Evolution of Subnational Regimes

36

Authoritarian regimes

Democratic regimes

Formosa

Cuba S. Arabia

Hybrid regimes

Non-electoral regimes

Mendoza Canada Electoral regimes

Uruguay

Measurements for Formosa Measurements for Mendoza

Figure  1.2.  A graphical representation of national and subnational regime variance and its measurement

indicators) of democracy for Formosa and Mendoza, respectively. Notice that the scores for each province differ considerably – because measurement error is inevitable, and likely not small given the aforementioned measurement difficulties – especially for the case of Formosa because its hybrid nature makes it harder to measure. In fact, the most democratic circle is not far from the least democratic triangle. In practical terms, this means that measurement has to be more careful and rigorous in the empirical domain of this book than at the national level. The scoring of a single indicator may be enough to accurately estimate the magnitude of the difference between Saudi Arabia and Uruguay. As the figure shows, the same is not true for Formosa and Mendoza: picking any pair of triangle and circle may lead to greatly under- or overestimating the regime differences between them. Given the complexities of a “thick” concept such as democracy, and the additional difficulties posed by subnational regimes (as compared to national regimes), I tackle descriptive inference applying to the Argentine provinces both objective and subjective strategies that have been used in political science to measure national democracy (Bollen and Paxton 2000, 60). The first approach measures democracy through objective, typically electoral and institutional indicators (well-­ known examples of objective indices of regime type are those of Alvarez et al. 1996 and Vanhanen 2000). These indicators are generally available from the historical record of many polities and years, and can usually be collected at a low cost in terms of time and money. An additional key advantage is that they involve little interpretation by the researcher (Vanhanen 2000, 255–7), which makes them highly reliable. Most of the time two independent coders will come to the same conclusion about whether a polity has multiparty elections or not, or what the vote share of the largest party was. The main disadvantages of objective indicators are that they tend to capture only one or a few “thin” (Coppedge 1999) aspects of democracy, and to do so with middling levels of validity. Objective measures typically focus on the Schumpeterian “competitive elections” aspect of democracy (or the “contestation” subdimension in Table 1.1),

Defining and Measuring Subnational Regimes

37

leaving out key dimensions such as checks and balances (“institutional constraints” in my operationalization), or respect for civil rights (“liberal rights” in Table  1.1). As for validity, many objective indicators (such as electoral outcomes) partly reflect factors other than underlying levels of democracy. Whether or not there is rotation in power or the margin by which an incumbent wins an election may say something, not only about how democratic the system is and about the extent to which there is a level electoral playing field, but also about such factors as the performance of an administration and the electoral rules (Bogaards 2007). The second tradition uses subjective or “perceptions-­based” (Kaufmann, Kraay, and Mastruzzi 2005) indicators.19 In this strategy, a researcher makes an informed judgment about a certain aspect of democracy in a given polity using secondary sources and/or consulting experts. Subjective operationalizations are behind well-­respected and widely used datasets in many subfields of political science.20 In fact, most mainstream national measures of democracy are based on subjective indicators (Bollen and Paxton 2000), including Polity IV (Marshall, Gurr, and Jaggers 2014), Freedom House’s (2016) ratings of political rights and civil liberties, Coppedge and Reinecke’s Polyarchy Index (1991), Mainwaring, Brinks, and Pérez-­Liñán’s (2001) index for Latin American countries, and most of the Varieties of Democracy indices (Coppedge and Gerring, et al. 2011).21 In the subjective strategy it is easier to tailor indicators so that they are valid measures of the theoretical concept of interest. Instead of relying on whether there is actual rotation in power as a proxy for the extent to which elections are free and fair, which inevitably leads to miscoding some democracies as autocracies (Alvarez et al. 1996), scholars can consult secondary sources and experts to assess exactly what they need to know – i.e., the level of electoral fairness and freedom. Moreover, in this strategy the researcher need not limit her- or himself to indicators that tap contestation only, as she or he can develop measures for all the dimensions and subdimensions of democracy contained in her or his conceptualization. Although the subjective approach is in principle more valid and thicker than the objective one, it is typically less reliable as different sources and experts may differ in their assessments of a given aspect of democracy in a given polity. This problem is compounded by the fact that perceptions-­based indices are often vague regarding the procedures to convert information from  The word “subjective” is often loaded with negative connotations. It is sometimes associated with normative biases or interested opinions. I use it in a straightforward, neutral way to describe a measurement process based on the informed and educated judgments of certain “subjects” (Mainwaring, Brinks and Pérez-­Liñán 2001, 38). 20  For example, the ideological positions of European parties have been estimated on the basis of several expert surveys (Hooghe et al. 2010). 21  Bowman, Lehoucq, and Mahoney (2005, 940) make a strong case in favor of measures of democracy based on the judgments of experts with deep knowledge of the polities to be coded. 19

38

Description: The Anatomy and Evolution of Subnational Regimes

secondary sources or experts into ratings. The Polity manual (Marshall, Gurr, and Jaggers 2014), for example, provides some guidelines to code the indicator “Constraints on Chief Executive,” but it is easy to see how different coders may disagree about whether a given case should be coded as “substantial limitations,” “intermediate category,” or “slight to moderate limitations.”22 Further inter-­coder reliability analyses revealed some significant disagreements among independent coders (Marshall, Gurr, and Jaggers 2014, 5–8). Likewise, Freedom House (2016) codes questions such as “are the electoral laws and framework fair?” on the basis of a number of subquestions that “are intended to provide guidance to the analysts regarding what issues are meant to be considered in scoring each checklist question.” Different coders, however, may draw on different sources or give more or less weight to different subquestions. As Freedom House (2016) indicates, “an element of subjectivity is unavoidable in such an enterprise,” but such an element is more troubling when the methodology provides little information about the specific sources used, the number of coders per country, and the levels of inter-­coder reliability. These shortcomings, however, can be improved taking a few, relatively simple steps such as listing the consulted sources, spelling out clear coding rules, having more than one person code each polity, and calculating and reporting measures of inter-­coder reliability. This means that it is possible to benefit from the higher validity of subjective measures without having to pay a high price in terms of reliability, as illustrated by the SEPP introduced here. Table 1.2 summarizes the advantages and disadvantages of the objective and subjective measurement strategies. The complementarity is clear: one tends to be strong where the other is weak. I therefore use both approaches to describe Argentina’s provincial regimes. The operationalization decisions and the details of the methodological design of each of the two measurement approaches are explained in Chapters 2 (the objective Subnational Democracy Index) and 3 (the subjective indices derived from the SEPP)23 so that any researcher can replicate the study – with Table 1.2.  Advantages and disadvantages of objective and subjective indicators of democracy Criteria

Objective indicators

Subjective indicators

Secondary data availability Cost of producing indicators Reliability Validity Conceptual coverage (“thickness”)

Usually high Low/moderate High Middling Typically low

Usually low Moderate/High Middling High Potentially very high

 These are three of the indicator’s seven categories.  See also Gervasoni (2010b) and Gervasoni (2016a).

22 23

Defining and Measuring Subnational Regimes

39

some adaptation to local political context – in any other country with elected and reasonably autonomous regional governments (the SEPP has already been replicated in Mexico, see Gervasoni, Loza and Méndez 2016). Hopefully, in the not-so-distant future, scholars will periodically produce estimates of regional regimes around the world which can be used to assess the causes and consequences of subnational democracy.

1.4  Aggregation: From Indicators to Indices of Subnational Regimes Aggregation  – the procedure for combining multiple measures of the same underlying variable into an index – is often the weakest part of operationalization efforts in the social sciences. Rules of aggregation are seldom clearly justified and sometimes not even spelled out. This is actually the case for many and important national-­level indices of democracy (Munck and Verkuilen 2002; Munck 2009). I provide a detailed explanation and justification of the aggregation rules for all the indices introduced in the following chapters (see Sections 2.1, 3.1.1, and 7.2). I anticipate the main procedures here. The objective Subnational Democracy Index is a factor score, i.e., a weighted average of five correlated indicators, expressed as a standardized variable. The subjective indices derived from the SEPP are unweighted averages of conceptually related and/ or highly correlated (and normalized) survey items. The objective Comparative Subnational Democracy Index introduced in Chapter 7 is the quadratic mean of three (normalized) indicators. Aggregation choices imply a theory of the concept, that is, clear ideas on how its different elements relate to each other (Goertz 2006: Munck 2009). The critical decision in combining our indicators into an index is about substitutability, or the extent to which high levels on one indicator can compensate for low levels on other indicators. For example, using the maximum operator – an aggregation rule in which the index takes the value of the indicator with the highest score  – implies full substitutability (Goertz 2006, 136): a polity with a single measure pointing in a democratic direction would be considered as democratic as another one with all indicators pointing in that direction. Conversely, the minimum operator implies no substitutability, as the worst-­ performing indicator determines the score of the index: a polity that does poorly on a single indicator would be considered as authoritarian as one doing poorly on all of them. Variants of the mean (such as the arithmetic, geometric, and quadratic means), on the other hand, imply partial substitutability: a low score on an indicator can be partially compensated by a high score on other indicators. Indicators that are not substitutable are akin to necessary conditions. Given the conceptualization of democracy I offer, the only non-­substitutable element of democracy is multiparty elections for the main government

40

Description: The Anatomy and Evolution of Subnational Regimes

positions. This bare minimum is present in the empirical domain of this book: Argentina and the seven countries I analyze in Chapter 7 have electoral regimes in all of their first-­level subnational units.24 Therefore, some substitutability is present in all indices I propose. Full substitutability, however, is also incompatible with my conceptualization: there is no single element of democracy that can compensate for weaknesses in other elements. Even electoral rotation in the executive cannot make up for human rights violations or government censorship of the media. Therefore, all the indices I propose implement partial substitutability using some kind of (unweighted or weighted, arithmetic or quadratic) mean. I provide the details and justification of the aggregation rules for each index in Chapters 2 (Subnational Democracy Index), 3 (SEPP-­based indices), and 7 (Comparative Subnational Democracy Index).

1.5 Conclusion This chapter presented the main operationalization decisions with respect to the dependent variable in this book  – the degree of subnational democracy. Operationalization implies three “challenges” (Munck and Verkuilen 2002): conceptualization, measurement, and aggregations. First, I defined democracy and identified its dimensions, subdimensions, components, and subcomponents. One important conceptual decision was to focus on the level of (subnational) democracy, not the quality of (subnational) democracy, a fuzzier concept. I also made it clear that in the context of national democratic regimes (such as those of Argentina and other countries analyzed in this book) the least democratic subnational polities are typically not authoritarian, but “hybrid”: complex mixtures of democratic and authoritarian elements. Second, I argued that measuring subnational democracy is especially challenging, and consequently introduced and justified two different measurement strategies, one objective and one subjective. Finally, I discussed alternative aggregation rules, that is, different ways of combining indicators of the same concept to form indices. The next two chapters present the details and results of each of the two measurement strategies. Chapter  2 focuses on the objective Subnational Democracy Index, a measure based on five electoral and institutional indicators. Chapter 3 does the same with the subjective indices of different aspects of subnational regimes based on data from the SEPP. The first sections of these chapters provide the details of the indicators involved, elaborate on their validity and reliability, and explain and justify the specific aggregation rules used to produce the indices. The rest of Chapters 2 and 3 use these indices to provide a comprehensive description of Argentina’s provincial regimes, of the differences among them, and of their evolution over time. 24

 Except in the rare cases when the national government removes local authorities and rules a region directly, as in Argentina’s federal interventions (see above note 11) or India’s presidential rule.

Defining and Measuring Subnational Regimes

41

The use of two different measurement strategies with complementary strengths and weaknesses allows for a description of subnational regimes of unprecedented breadth and depth. In fact, the next two chapters are, to my knowledge, the first effort to assess the level of democracy in all the subnational units of a country using both objective and subjective measures. Descriptive inferences contain an inevitable element of uncertainty, but when alternative measurement strategies arrive at the same conclusion – as it is the case for several Argentine provinces – our confidence in those inferences grows significantly.

2 The Subnational Democracy Index Trends in Provincial Regimes (1983–2015)*

This chapter focuses on the objective Subnational Democracy Index (SDI). This index taps two core dimensions of democracy – contestation and power concentration in the incumbent. The electoral and institutional indicators that measure these dimensions, however, are empirically unidimensional. This unidimensionality has the advantage of simplifying the description of subnational regimes: a single summary measure provides an overall characterization of all provinces in all gubernatorial periods from 1983 (the year in which democracy was reestablished in Argentina) to 2015 (the year of the last gubernatorial election). A more comprehensive, multidimensional (and more valid) description of subnational democracy in Argentina is presented in Chapter 3 (dedicated to the Survey of Experts on Provincial Politics [SEPP]). That description, however, lacks the temporal component of the SDI, as the survey only measured the characteristics of provincial regimes for the 2003–7 period. The first section of this chapter discusses the methodological design of the SDI. The second sections analyzes its results, both in terms of differences among provinces and of trends over time. The final section complements the SDI’s quantitative results with qualitative evidence describing the undemocratic institutions and practices that differentiate hybrid provinces from the more democratic ones.

2.1  Objective Indicators: The Subnational Democracy Index Although the most popular indices of democracy at the national level, such as Polity IV and Freedom House, are subjective, the opposite is the case for *

  Portions of this chapter were published previously in Gervasoni (2010a).

42

The Subnational Democracy Index

43

the literature on subnational regimes. Probably because objective indicators on the existence and results of regional elections around the world are relatively easy to obtain, the few scholars who have attempted to systematically measure subnational democracy in a given country have typically used objective measures (Hill 1994; Hernández Valdez 2000; Beer and Mitchell 2006; Goldberg, Wibbels, and Mvukiyehe 2008; Gervasoni 2010a; Giraudy 2010, 2015; Borges 2016). As explained in Section 1.2.1, the least democratic subnational units within national-­level democracies – in Argentina and elsewhere – are better characterized as hybrid rather than authoritarian. The rulers of autonomous (but not sovereign) regional hybrid regimes are vulnerable to the threats of national political actors, from the media and public opinion to federal authorities. In this context, they have incentives to showcase a democratic institutional architecture, avoid openly authoritarian rules or practices, and rely exclusively on stealthy tactics to restrict political rights. Elections are held and ballots counted fairly, but incumbents massively outspend challengers; the local media is formally independent, but it is bought off to bias coverage in favor of the ruling party; dissidents are not jailed, just excluded from coveted public jobs. This combination of democratic-­looking electoral institutions with subtle and hidden violations of democratic principles makes hybrid regimes difficult to measure (see Section 1.3). A practical alternative is to focus on the effects of those violations on political outcomes, which are more amenable to empirical observation than the violations themselves. Scholars of regimes have generally (and reasonably) assumed that authoritarian practices within an electoral regime increase the probability of extended stints of single-­party rule, unusually large electoral majorities for incumbents, overwhelming executive control of the legislature, no term limits, and so forth. None of these political outcomes defines the type of regime, but following the logic of “effect indicators” (Bollen and Lennox 1991), they should reflect changes in the underlying level of democracy. These indicators are correlated with the trait of interest because they are an effect of it (in the same way that physicians use symptoms typically caused by a disease as indicators of its presence). Scholars who state that “[d]emocracy is a system in which [incumbent] parties lose elections” (Przeworski 1991, 10), that “no country in which a party wins 60 percent of the vote twice in a row is a democracy” (Przeworski 1991, 95), and that “[c]ountries in which one party wins an overwhelming share of seats are not likely to be democracies” (Alvarez et  al. 1996, 13) know that democracy is not about any specific electoral outcome. They also know, however, that certain outcomes are typical effects of its weakness or absence. If we agree with them that democracy requires “an opposition that has some chance of winning office as a consequence of elections” (Alvarez et al. 1996, 5), then provinces like Formosa, San Luis or Santa Cruz, where the same party has controlled the governorship for nine consecutive four-­year terms since 1983 – often obtaining over 2/3 of the votes  – are less democratic than

44

Description: The Anatomy and Evolution of Subnational Regimes

Entre Ríos or Mendoza, where different parties have alternated in office, and where incumbents have seldom gained more than 50 percent of the votes. Using levels of electoral contestation to operationalize democracy can also be justified because they are “causal indicators” (Bollen and Lennox 1991), that is, measures that correlate with the underlying trait of interest not because they are an effect of it, but because they are a cause of it. There is evidence that vigorous electoral competition leads, causally, to higher national and subnational government responsiveness (Besley and Burgess 2002; Griffin 2006; Hobolt and Klemmensen 2008), and, in the case of the Argentine provinces, to democracy-­enhancing independent judiciaries (Chavez 2003, 2004). If observable levels of electoral contestation are both a cause and an effect of underlying levels of democracy, then the former should be a reasonably valid indicator or the latter. Some national indices of regime type include indicators of electoral competitiveness. Dahl’s measures of polyarchy (1971, 238), for example, differentiate between competitive and partially competitive regimes using a threshold of 85 percent of legislative seats controlled by a single party. An explanation of Freedom House’s indices states that “[t]he extent of democratic rights can also be empirically suggested by the size of the opposition vote. While on rare occasions a governing party or individual may receive overwhelming support at the polls, any groups or leader that regularly receives seventy percent or more of the vote indicates a weak opposition and the probable existence of undemocratic barriers in the way of its further success” (Gastil 1991, 29). Vanhanen (2000, 257) explicitly agrees with this 70 percent threshold, and one of the two measures of his index is a function of the percentage of votes won by the largest party. Notice, however, that the authors cited above disagree in a very important respect: while some view any large electoral or legislative majority as less democratic, others only doubt democracy is fully at work when incumbents obtain such majorities. Both national and subnational indices differ on this key aspect. Vanhanen’s (2000) contestation penalizes a large proportion of the vote for any party, incumbent or challenger, and so do many of the existing subnational measures of regime type, for example the adaptation of Vanhanen’s index used by Beer and Mitchell, the various versions of the effective number of parties (ENP) (e.g., Hernández Valdez 2000; Giraudy 2010), or the margin of victory in gubernatorial elections (Goldberg, Wibbels, and Mvukiyehe 2008). On the other hand, the dichotomous index developed by Alvarez et al. (1996) rewards opposition landslides, as they constitute solid evidence that incumbent “parties lose elections.” Examples of this approach at the subnational level are measures such as the vote share of the incumbent governor’s party (used by Borges 2007; Goldberg, Wibbels, and Mvukiyehe 2008; Gervasoni 2010a) or the various measures of the incumbent’s legislative strength (part of the indices designed by Borges 2007; Gervasoni 2010a; Giraudy 2010). The SDI I introduce below adopts the “incumbent” criterion, not the “largest party” criterion. It considers, “in the spirit of Alvarez et al. (1996), that a

The Subnational Democracy Index

45

handily defeated incumbent is an expression of healthy levels of democracy” (Gervasoni 2010a, 317, fn. 96). In practice, the distinction may be often irrelevant, as large electoral or legislative majorities are obtained much more frequently by incumbents than by challengers.1 Conceptually the distinction is critical, and it may be empirically critical too in contexts in which strong showings of the opposition are more than an unusual quirk. The SDI focuses on two key dimensions of democracy (out of the five identified in Section 1.2.2): contestation, which is central to all definitions (Dahl 1971; Alvarez et al. 1996), and power concentration in the incumbent (or institutional constraints on the power of the government), a critical component of liberal or “protective” understandings of democracy (Held 1987). Other important aspects of democracy, such as liberal rights or effective elections, cannot be incorporated because of the unavailability of comparable objective data for all provinces. The index includes three indicators of (electoral) contestation – Executive contestation, Legislative contestation, and Succession control – and two indicators of power concentration in the incumbent – Legislature control and Term limits.2 Measures of electoral contestation have been shown to be highly correlated with mainstream subjective indices of political rights and freedoms (Vanhanen 2000, 256; Coppedge, Alvarez, and Maldonado 2008). For example, for the period 1970–2000 Vanhanen’s contestation indicator (100 minus the percentage of total votes cast won by the largest party) correlates at 0.89 with the Polity Index, at 0.86 with Freedom House’s Political Rights and at 0.87 with V-­Dem’s Electoral Democracy Index.3 It should be encouraging for our purposes that a single electoral contestation indicator, which does not even, distinguish whether the largest party is the incumbent or not, can quite accurately predict the level of democracy as measured by standard indices. A measure such as the SDI, that combines several indicators of this type and that focuses on the performance of the incumbent (not the largest party), should perform even better. In the following paragraphs, I explain and justify the five indicators that constitute the SDI. 1. Executive contestation measures the extent to which there are real chances for the opposition to defeat the governor’s party. It is simply the proportion of the valid vote won by the incumbent

 For example, Saikkonen (2016a), who uses the “largest party” criterion in her study of Russia’s regions, indicates that 94.5 percent of “hegemonic” winners (parties wining the governorship with more than 70 percent of the vote) in her sample are incumbents (p. 270). 2  Vote and seats figures are from the Dirección Nacional Electoral and the Atlas Electoral de Andy Tow (at www.mininterior.gov.ar/elecciones, and http://andy.towsa.com, respectively). 3  My calculations using the Varieties of Democracy dataset “V-­Dem [Country-Year/Country-Date] Dataset v6.2” (Coppedge et al. 2016). All p-­values